Safe withdrawal ideas, thoughts and provocations

Financial Independence/Retire Early -- Learn How!
hocus
Moderator
Posts: 435
Joined: Mon Dec 02, 2002 12:56 am

Post by hocus »

I would say it is not a conservative viewpoint, simply a correct one....I think it is an important distinction.

I agree.

When you are purporting to rely on data to help you analyze a question, you must look at all the data that bears on that question to have any hope of generating an accurate assessment of it. It is not "conservative" to do so, or "liberal" not to do so. It is using a valid methodology to do so and using an invalid methodology not to do so.
hocus
Moderator
Posts: 435
Joined: Mon Dec 02, 2002 12:56 am

Post by hocus »

The end result is the same either way.

I agree.

The practical problem that we are trying to address is that the conventional methodology does not provide a reaonsable assessment of what the historical data reveals to be safe. If aspiring early retirees place their confidence in studies done using the conventional methodology, there is a good chance that they will suffer serious life setbacks.

Any approach that protects people from those serious life setbacks does the job that needs to be done. It appears to me that Petey's approach and my approach would both prodce the same practical results.
User avatar
ataloss
**** Heavy Hitter
Posts: 559
Joined: Mon Nov 25, 2002 3:00 am

Post by ataloss »

However I will point to the graph at the bottom of this page on REHP and note that 'average' valuations have also produced actual hSWRs in the vicinity of 4%.


which is why I am not comfortable with the discussion of a "swr" substantially greater than 4% based on undervaluation.

Gummy had posted something about maximal retrospective 30 year withdrawal rates which were as high as 10% IIRC but to be "s" I think a withdrawal rate needs to be more in line with the historical record :wink::wink:
Have fun.

Ataloss
peteyperson
**** Heavy Hitter
Posts: 525
Joined: Tue Nov 26, 2002 6:46 am

Re: careful

Post by peteyperson »

I do agree there, Ben.

I think though that it is wrong to hide the new valuation in the calculation of the swr. You should first reach the right valuation. Then later determine a safe withdrawal rate. That calculation is completely seperate and is unaffected by whether the asset base is $100k or $1m. That's why I think it should be seperate.

Also, if correcting the misunderstanding over true values is a key component to steering people in the right direction from the beginning (which it would seem to need to be given the stories I've highlighted where valuation can lead people down the wrong path), then it should be seperate to highlight its own importance.

So I agree the result (if we agreed on the swr calculation methodology & data) would be the same, but the seperation is a priority I believe. I have a suspicion that hocus may change his opinion on this point. Hocus wants the swr to be seperate from other considerations, to be a mathematic construct. Its a wrongheaded approach to retro-fit the valuation on the same spreadsheet as you work out the swr. It will lead to lower w/d rates possibly and will create questions. Seperating the two allows for less possible confusion, simplicity and clarify. On a website, one page on "Valuation, how to work out what you're really worth" and a second on how we calculate SWR. A better result surely.

Let me know what you think.

Petey


BenSolar wrote:
peteyperson wrote: I don't know if I believe this to be true at all.

If the valuation of your portfolio is re-calculated to take full account of overvaluation, then how does overvaluation matter? ...
hocus wrote: Changes in valuation levels always affect SWRs.


Let's be careful that we don't start talking past each other. Hocus, IINM, was speaking from his traditional viewpoint of adjusting SWR, not the valuation. The end result is the same either way.

Ben
peteyperson
**** Heavy Hitter
Posts: 525
Joined: Tue Nov 26, 2002 6:46 am

Post by peteyperson »

I concur. My point exactly.

Petey

hocus wrote: I would say it is not a conservative viewpoint, simply a correct one....I think it is an important distinction.

I agree.

When you are purporting to rely on data to help you analyze a question, you must look at all the data that bears on that question to have any hope of generating an accurate assessment of it. It is not "conservative" to do so, or "liberal" not to do so. It is using a valid methodology to do so and using an invalid methodology not to do so.
peteyperson
**** Heavy Hitter
Posts: 525
Joined: Tue Nov 26, 2002 6:46 am

Post by peteyperson »

It would produce the same end result, but it would seperate out corrected valuation from safe withdrawal calculation. As I mention to Ben, this has considerable benefits as most people are presumably out there working from wrong valuations (basically whatever number is detailed in their brokerage account).

It is a fundamental error for them. If the result is the same, shouldn't our approach be for clarity? If so, wouldn't seperating each issue out, first getting the right valuation and later performing a swr calculation based on whatever figure is punched in, wouldn't that be clearer? This might not be possible if you can only come up with a swr calculation that works when including valuation elements. Bogle always stresses simplicity over complexity. Simple presentation is likely to be better understood. Valuation correction, this is a night and day issue. It has to be fixed first.

Please see my reply to Ben Solar for further clarity.

Petey


hocus wrote: The end result is the same either way.

I agree.

The practical problem that we are trying to address is that the conventional methodology does not provide a reaonsable assessment of what the historical data reveals to be safe. If aspiring early retirees place their confidence in studies done using the conventional methodology, there is a good chance that they will suffer serious life setbacks.

Any approach that protects people from those serious life setbacks does the job that needs to be done. It appears to me that Petey's approach and my approach would both prodce the same practical results.
User avatar
BenSolar
*** Veteran
Posts: 242
Joined: Mon Nov 25, 2002 5:46 am
Location: Western NC

Re: careful

Post by BenSolar »

peteyperson wrote: Seperating the two allows for less possible confusion, simplicity and clarify. On a website, one page on "Valuation, how to work out what you're really worth" and a second on how we calculate SWR. A better result surely.


I like the idea.
"Do not spoil what you have by desiring what you have not; remember that what you now have was once among the things only hoped for." - Epicurus
hocus
Moderator
Posts: 435
Joined: Mon Dec 02, 2002 12:56 am

Post by hocus »

Technically the valuation is "real" at any given moment in the sense that you could choose to sell all your shares and they would get the best price on the open market at that time.

That's right. Valuing your portfolio pursuant to the numbers reported in the newspaper gives you a number that is real in the short-term but not real in the long term. SWR analysis purports to tell you what is safe for the long-term, but it uses a number that is real only in the short-term to do so. That doesn't work.

I've yet to grasp how starting from an overvalued market, your w/d rate goes down.

Here's an example that uses rough numbers to illustrate the concept.

Say that the value of your portfolio is $1,000,000 according to the numbers in that morning's newspaper. Say that stocks are currently at price levels double the historical norm.

Your approach says that you must adjust the porfolio amount down to $500,000, and then you can take 4 percent of that as your withdrawal--that's $20,000.

With my approach, you value the portfolio at $1,000,000, but you take a withdrawal of only 2 percent--that's $20,000.

The two approaches produce the same results in practical terms. Both are addressing the need to consider valuation in determing how much you can take from your portfolio each year and have a plan that the historical data indciates is safe.
hocus
Moderator
Posts: 435
Joined: Mon Dec 02, 2002 12:56 am

Post by hocus »

to be "s" [safe], I think a withdrawal rate needs to be more in line with the historical record

I'm saying that the historical record you look at must be the relevant historical record for the question you are trying to answer.

The conventional methodology uses the same historical record to assess safety for retirements beginning in times of low valuation, moderate valuation, and high valuation. That's an invalid approach because it ignores the effects of changing valuation levels on returns expectations. It is not reasonable to expect the same sorts of returns going forward from a retirement year of high valuation as it is to expect going forward from a year of low valuation.

If you are retiring in a year of low valuation, you need to look at what happened to retirements starting from other years of low valuation. If you do that, I am confident that you will find that a withdrawal rate of higher than 4 percent is revealed to be safe.
peteyperson
**** Heavy Hitter
Posts: 525
Joined: Tue Nov 26, 2002 6:46 am

Post by peteyperson »

Hi hocus,

To appreciate your point, I would need to understand what calculation has been done to generate the 2% number from the inflated figures, and whether the same calculation could be applied on figures less inflated or even undervalued.

I certainly apprecaite your point that the end figure is the same.

Petey


hocus wrote: Technically the valuation is "real" at any given moment in the sense that you could choose to sell all your shares and they would get the best price on the open market at that time.

That's right. Valuing your portfolio pursuant to the numbers reported in the newspaper gives you a number that is real in the short-term but not real in the long term. SWR analysis purports to tell you what is safe for the long-term, but it uses a number that is real only in the short-term to do so. That doesn't work.

I've yet to grasp how starting from an overvalued market, your w/d rate goes down.

Here's an example that uses rough numbers to illustrate the concept.

Say that the value of your portfolio is $1,000,000 according to the numbers in that morning's newspaper. Say that stocks are currently at price levels double the historical norm.

Your approach says that you must adjust the porfolio amount down to $500,000, and then you can take 4 percent of that as your withdrawal--that's $20,000.

With my approach, you value the portfolio at $1,000,000, but you take a withdrawal of only 2 percent--that's $20,000.

The two approaches produce the same results in practical terms. Both are addressing the need to consider valuation in determing how much you can take from your portfolio each year and have a plan that the historical data indciates is safe.
peteyperson
**** Heavy Hitter
Posts: 525
Joined: Tue Nov 26, 2002 6:46 am

Post by peteyperson »

Perhaps this is true but only where you are willing to move your w/d rate up and down depending on the valuation. That approach isn't practical for most people trying to set a budget and know what their target sum is (valuation adjusted or not).

In other words, it is no good being 2% at one point in history but 4% in other times. You cannot slash your budget in half each time that happens, nor can you time your retirement to only be in a low or high valuations that day. It makes the whole thing a moving target and I reject that approach. You do have to make set assumptions of overall performance on the selected asset classes during Distribution phase, but that is intended to smooth out the girations and deliver a measured reliable income stream. The idea you are discussing fails to do that.

Hocus, would you be able to explain the lower w/d rate in more depth, referring perhaps to my own statements about using a large cash buffer to smooth out returns making market fluctuations unimportant unless they go multi-decade? What data do well have to show from markets that are high, what real return per year was delivered over the decade following? That probably goes to the heart of the issue. I'm coming at it from the perspective that the worst period was 15 years where we rose back to the previous value inc. inflation over that 15 year period. No growth in real terms which would have depleted the overall value of the portfolio even if you had enough cash and didn't have to sell i.e. there wasn't the 2%-4% real growth expected in that asset class.

I'm going to read Shiller's report on this issue to understand his perspective on how valuations affect things. I'm importing the Four Pillars book from the States and will have digested that in about three weeks. I hope to then come back with more thoughts at that time. I will either be agreeing with you and we shall don lightsabers to the death! :P I shall enlist Ben Solar on cheerleader support and you can have a choice of the esteemed Wanderer or the equally masterful John R. :o

Petey


hocus wrote: to be "s" [safe], I think a withdrawal rate needs to be more in line with the historical record

I'm saying that the historical record you look at must be the relevant historical record for the question you are trying to answer.

The conventional methodology uses the same historical record to assess safety for retirements beginning in times of low valuation, moderate valuation, and high valuation. That's an invalid approach because it ignores the effects of changing valuation levels on returns expectations. It is not reasonable to expect the same sorts of returns going forward from a retirement year of high valuation as it is to expect going forward from a year of low valuation.

If you are retiring in a year of low valuation, you need to look at what happened to retirements starting from other years of low valuation. If you do that, I am confident that you will find that a withdrawal rate of higher than 4 percent is revealed to be safe.
hocus
Moderator
Posts: 435
Joined: Mon Dec 02, 2002 12:56 am

Post by hocus »

It is no good being 2% at one point in history but 4% in other times.... You cannot slash your budget in half each time that happens, nor can you time your retirement to only be in a low or high valuations that day. It makes the whole thing a moving target and I reject that approach.

You don't need to adjust your withdrawal rate each time valuations go up or down. You need to incorporate the valuation effect at the time you assess what is safe, which is generally the date you retire..

Presuming that you did the calculation properly then, there should be no need to change it from that point forward. If you did the calculation properly, and if the future turns out like the past, you will end up OK. If the future does not turn out like the past, then all bets are off. That is a caveat of all SWR analysis.

Nor do you need to "time" your retirement. When you use a valid SWR methodology, there is no advantage to retiring at a time of high valuation or low valuation. At a time of high valuation, the dollar value of your portfolio will be up, but the SWR will be down. At a time of low valuation, the dollar value of your portfolio will be down, but the SWR will be up. The one effect will cancel out the other.

To me, this cancelling-out effect is evidence of the superiority of an SWR methodology that considers the effect of changes in valuation. When people use the conventional methodology, you always get questions like "if stock prices drop dramtically in the year before I planned to retire, can I go back and use the value they were at in an earlier year as the number on which to calculate my safe withdrawal?"
This question came up several times on the REHP board in the year 2000.. Advocates of the conventional methodology get themselves caught up in all sorts of silly arguments trying to argue why it is OK or not OK to do this.

The answer is--if you calculated the SWR accurately, it won't make any difference whether valuation levels go up or down just before or after you retire. All that should count are the profits generated by the businesses you hold. Price level changes that are purely adjustments to the "fluff" factor should make no difference, and, when you incorporate the valuation factor into the SWR analysis, they do not.

If you calculate the SWR accurately, a big drop in prices in the year you plan to retire causes you no problem. Your portfolio value goes down, but the SWR goes up. The value of the underlying business assets you own did not change just because price levels changed. When you do not account for valuation in the analysis, all the results produced are messed up by the "noise" factor of rises or drops in the fluff factor. Incorporate the effect of valuation changes into the analysis, and this factor is properly excluded from the results generated by the analysis.

A valid SWR calculation measures the thing you really want to measure--the long-term value of the assets you are using to finance your retirement. The conventional methodology measures that plus a lot of fluff, and the inclusion of fluff in the calculations throws the results off from what the historical data indicates is safe.. My proposal is to de-fluffify SWR calcuations.
User avatar
ataloss
**** Heavy Hitter
Posts: 559
Joined: Mon Nov 25, 2002 3:00 am

Post by ataloss »

To appreciate your point, I would need to understand what calculation has been done to generate the 2% number from the inflated figures
Have fun.

Ataloss
hocus
Moderator
Posts: 435
Joined: Mon Dec 02, 2002 12:56 am

Post by hocus »

To appreciate your point, I would need to understand what calculation has been done to generate the 2% number from the inflated figures

Ataloss:

I have not seen the work papers that William Bernstein used to determine that the SWR for a high-stock portfolio at the top of the bull market was 2 percent. But I trust that he did not pull a rabbit out of a hat to come up with the number. My strong belief is that the number he reported on Page 234 of his book "The Four Pillars of Investing" was based on a fair and reasoned analysis of the data bearing on the question we are trying to answer.

Do you have any reason to think different?

In the event that he pulled that number out of a hat, you will get a chance to ask him about it in the event that he accepts my invitation to serve as one of our future "A Night With the NFB Group" guests. I am confident that you are going to be reassured from that conversation that William Bernstein is not the sort of individual who determines SWRs by pulling them out of a hat.
peteyperson
**** Heavy Hitter
Posts: 525
Joined: Tue Nov 26, 2002 6:46 am

Post by peteyperson »

Well you certainly win the award this month for the most inventive English!
hocus wrote: My proposal is to de-fluffify SWR calculations.



See, the way you do it, I find that confusing. The rate moves about depending on the valuation at the time of FIRE. When you adjust the portfolio back down to intrinsic value, your rate stays the same regardless of the current price. That's much clearer. Having the rate rise when valuation is below P/E initially confuses the hell outta me. If you retire and your assets are priced at P/E of 10, I would have to consider that the value of my assets. I wouldn't rachet them up to the mean 14.1. That may seem contradictory but I would rather be more cautious and not come unstuck later if we have 2-3 decades of 1970s style valuations that don't come back in a meaningful way. In that sense, I would rather prefer to date the accountant's girlfriend, Prudence.

I do fundamentally understand that a 28.2 P/E portfolio would have a 2% w/d rate and if the average P/E is 14.1, it would deliver a 4% w/d rate. I understand that a $100,000 portfolio based on intrinsic valuation but currently valued at $200,000 would need a 2% w/d rate, delivering $4,000 a year. I understand that a $100,000 portfolio of stocks sitting right on a 14.1x P/E valuation would be a 4% w/d rate, again delivering surprise! $4,000 a year. I was personally equating initially a different withdrawal rate with a lower $ amount. This I think is a natural reaction. I have to jump thru mental hoops to discover we're talking about the same amount of money ultimately and you're just fumaxing the calculation to account for the P/E variance.

I think that is a mentally complex way around the issue. You're effectively saying the w/d rate doesn't matter/is relative because it is dependent on the valuation and will come out the same on real $ terms either way. In my mind I have to do all sorts of twists and turns just to see that. I think many other people may too. Or maybe I'm just slow :)

It would be far simpler to accept a valuation under 14.1 as the current value and be cautious there, and re-value anything above 14.1 down to the historical mean. Then you use whatever the w/d rate is determined to be based on historical performance of the selected range of asset classes & their relative apportionment and apply it to the real value of your investments at that time. For example: Portfolio is $200,000 at 28.8 P/E. Re-value to $100,000. Asset mix is individual and will deliver a different w/d rate depending on the selection, the investments costs, possibly the timeline etc. You plug in the $100,000 adjusted true value figure and it spits out the w/d rate relative to the data you're entered. That's a simple, clear, clean, understandable approach. Messing with the w/d rate figure even though the w/d $ is the same is just a confusing way to deal with variable valuations to ultimately arrive at the same figure. I'm going for simplicity a la Bogle. Your way is more complicated, goes all round the houses mentally and doesn't deliver anything extra for that complexity! And the fact that the $ amount is actually the same is often not even mentioned or gets lost as it did here.

If you're saying it is also something to do with the fact that funding the Distribution phase from a high P/E has a history of causing funds to go bust because they deliver sub-par returns in the early years relative to your expected withdrawals, then that is something else entirely and we shouldn't confuse the two. I would then say you either come up to adjust w/d rate based on the P/E of stocks on the day of retirement or more helpfully perhaps, you come up with a w/d rate that works based on history & whether you're inflated upon retirement or not. A w/d rate that's good for all time zones, so to speak. But again, these are two different issues. Note: I don't think you're talking about this at all.

How is the w/d rate calculated and at the same time takes into account the valuation in the figure it generates. In other words, if you plug in $200,000 instead of $100,000, is there a field for the P/E of the holdings so that the w/d rate is readjusted using that or is it done some other way? In other words, how does the spreadsheet know you're overvalued? I'm assuming this is the way it knows.

I feel we are making good progress or at least I am understanding the way you are working this out a bit more. Keep going, please.

Petey


hocus wrote: It is no good being 2% at one point in history but 4% in other times.... You cannot slash your budget in half each time that happens, nor can you time your retirement to only be in a low or high valuations that day. It makes the whole thing a moving target and I reject that approach.

You don't need to adjust your withdrawal rate each time valuations go up or down. You need to incorporate the valuation effect at the time you assess what is safe, which is generally the date you retire..

Presuming that you did the calculation properly then, there should be no need to change it from that point forward. If you did the calculation properly, and if the future turns out like the past, you will end up OK. If the future does not turn out like the past, then all bets are off. That is a caveat of all SWR analysis.

Nor do you need to "time" your retirement. When you use a valid SWR methodology, there is no advantage to retiring at a time of high valuation or low valuation. At a time of high valuation, the dollar value of your portfolio will be up, but the SWR will be down. At a time of low valuation, the dollar value of your portfolio will be down, but the SWR will be up. The one effect will cancel out the other.

To me, this cancelling-out effect is evidence of the superiority of an SWR methodology that considers the effect of changes in valuation. When people use the conventional methodology, you always get questions like "if stock prices drop dramtically in the year before I planned to retire, can I go back and use the value they were at in an earlier year as the number on which to calculate my safe withdrawal?"
This question came up several times on the REHP board in the year 2000.. Advocates of the conventional methodology get themselves caught up in all sorts of silly arguments trying to argue why it is OK or not OK to do this.

The answer is--if you calculated the SWR accurately, it won't make any difference whether valuation levels go up or down just before or after you retire. All that should count are the profits generated by the businesses you hold. Price level changes that are purely adjustments to the "fluff" factor should make no difference, and, when you incorporate the valuation factor into the SWR analysis, they do not.

If you calculate the SWR accurately, a big drop in prices in the year you plan to retire causes you no problem. Your portfolio value goes down, but the SWR goes up. The value of the underlying business assets you own did not change just because price levels changed. When you do not account for valuation in the analysis, all the results produced are messed up by the "noise" factor of rises or drops in the fluff factor. Incorporate the effect of valuation changes into the analysis, and this factor is properly excluded from the results generated by the analysis.

A valid SWR calculation measures the thing you really want to measure--the long-term value of the assets you are using to finance your retirement. The conventional methodology measures that plus a lot of fluff, and the inclusion of fluff in the calculations throws the results off from what the historical data indicates is safe.. My proposal is to de-fluffify SWR calcuations.
Last edited by peteyperson on Tue Jul 15, 2003 12:32 pm, edited 1 time in total.
hocus
Moderator
Posts: 435
Joined: Mon Dec 02, 2002 12:56 am

Post by hocus »

Your way is more complicated and doesn't deliver anything extra for that complexity.

Perhaps that is so in a purely theoretical sense. In a practical sense, I don't think so.

If the adjustment for valuation were made in the way that you suggest, there would need to be language added to all SWR studies noting that the rate reported is valid only in cases in which the person using the information first makes adjustments in his portfolio value to reflect prevailing Price/Earnings levels.

Someone is one a radio show and the host says "So what is the SWR?" Does the person say "4 percent?" or "4 percent, but only if you adjust your portfolio down to relfect the current valuation level" or "2 percent for most people nowadays" or "it depends" or what?

I accept the validity of your approach on a theoretical level. I think it has some serious complexities of its own, however. Personally, I think that incorporating the valuation factor into the SWR calculation is the simpler approach. There is room for reasonable differences of opinion on this question, however.

That may seem contradictory but I would rather be more cautious and not come unstuck later if we have 2-3 decades of 1970s style valuations that don't come back in a meaningful way.

The SWR number is always an extremely conservstive number. It is the product of an analysis of the worst case scenario. I think that is conservatism enough.

If you really want to be more conservative that that in your own plan, it's your business. But the way to do that is to make an adjustment to the SWR to suit your own desires, not to change the way that the SWR itself is calculated.

If it is truw that a 5 percent withdrawal at a time of low valuation is every bit as safe as a 3 percent withdrawal at a time of overvaluation, according to the historical data, a valid SWR study should let people know this. It is up to the people who make use of the analysis how conservative they want to be. But the first step should be to report the true SWR accurately, and the historical data indicates that it is just as safe to take withdrawals above 4 percent for retirements beginning at times of high valuation as it is to take withdrawals below 4 percent for retirements beginning at times of high valuation.

Our role is to tell people what the data says, not to doctor the data in ways that support our own personal views as to how conservative or liberal we think people should be in their expectations of what the future will deliver.. Just go with the data and you can't go wrong, in my view. Leave it to each aspiring early retiree to decide for himself or herself how to adjust the number that the data produces to his or her personal circumstances.
peteyperson
**** Heavy Hitter
Posts: 525
Joined: Tue Nov 26, 2002 6:46 am

Post by peteyperson »

Sounds like you missed out something here. You're making the intercst mistakeTM. How would the radio host know if the portfolio was valued fairly or overvalued using a 4% withdrawal rate? You've said yourself, if the valuation is double the historical norm, the rate would halve from 4% to 2%. So if the caller couldn't answer the question of what is the average P/E of your stock market investments, how could the host know if their w/d rate was correct? So either way you'll need to get into the P/E associated with the portfolio value to know at what level of valuation is at in order to answer the w/d rate question correctly. This is using your methodology of muddling it all together. If you talk to the caller suggesting that they need to determine the true value first, sending them away to get that information and then call back with it. From that starting point, the answer becomes easy: 4% of the intinsic value of your assets (assuming it is determined that 4% is the safe w/d rate on their selection of assets allocated in the proportion so determined). It's not even a set 4% w/d rate answer even like that, the rate will change depending on the asset allocation, investment costs, the respective return on the investments overall and the current inflation rate. It will never be a set: " 4% is a safe withdrawal rate " unless you have a simple stock/cash allocation where the expected returns can be simply detailed.
hocus wrote: Someone is one a radio show and the host says "So what is the SWR?" Does the person say "4 percent?" or "4 percent, but only if you adjust your portfolio down to relfect the current valuation level" or "2 percent for most people nowadays" or "it depends" or what?


How does the spreadsheet account for the valuation assessment? Do you plug in the P/E figure along with the asset value?

Is there anything out there in spreadsheet land that has multiple fields like I suggested with different asset class investments, their costs and avg returns? Is it true that the swr figures calculated use commercial paper where you cannot get that return after costs and doesn't allow for some of the cash component to be REITS, for instance, that deliver a return higher than cash but lower than stocks? How do you account for the variety of investments people on the board are using now?

I suppose I am now distrustful of a safe withdrawal rate in an overly complicated calculation partly because I might not appreciate all the nuiances of how it is calculated (meaning I cannot take it as gospel), partly because I don't buy the intercst party line either. His 4% w/d rate good for all timezones clearly doesn't take account of inflated markets and as such is completely invalidated. Has intercst really not accepted this obvious truth once explained? Inflated prices would logically allow you to take a lower percentage of the total assets valued that day. I will have to get my head around how some of the swr are actually calculated in order to trust it.

Petey



hocus wrote: Your way is more complicated and doesn't deliver anything extra for that complexity.

Perhaps that is so in a purely theoretical sense. In a practical sense, I don't think so.

If the adjustment for valuation were made in the way that you suggest, there would need to be language added to all SWR studies noting that the rate reported is valid only in cases in which the person using the information first makes adjustments in his portfolio value to reflect prevailing Price/Earnings levels.

Someone is one a radio show and the host says "So what is the SWR?" Does the person say "4 percent?" or "4 percent, but only if you adjust your portfolio down to relfect the current valuation level" or "2 percent for most people nowadays" or "it depends" or what?

I accept the validity of your approach on a theoretical level. I think it has some serious complexities of its own, however. Personally, I think that incorporating the valuation factor into the SWR calculation is the simpler approach. There is room for reasonable differences of opinion on this question, however.

That may seem contradictory but I would rather be more cautious and not come unstuck later if we have 2-3 decades of 1970s style valuations that don't come back in a meaningful way.

The SWR number is always an extremely conservstive number. It is the product of an analysis of the worst case scenario. I think that is conservatism enough.

If you really want to be more conservative that that in your own plan, it's your business. But the way to do that is to make an adjustment to the SWR to suit your own desires, not to change the way that the SWR itself is calculated.

If it is truw that a 5 percent withdrawal at a time of low valuation is every bit as safe as a 3 percent withdrawal at a time of overvaluation, according to the historical data, a valid SWR study should let people know this. It is up to the people who make use of the analysis how conservative they want to be. But the first step should be to report the true SWR accurately, and the historical data indicates that it is just as safe to take withdrawals above 4 percent for retirements beginning at times of high valuation as it is to take withdrawals below 4 percent for retirements beginning at times of high valuation.

Our role is to tell people what the data says, not to doctor the data in ways that support our own personal views as to how conservative or liberal we think people should be in their expectations of what the future will deliver.. Just go with the data and you can't go wrong, in my view. Leave it to each aspiring early retiree to decide for himself or herself how to adjust the number that the data produces to his or her personal circumstances.
Another Peter
* Rookie
Posts: 1
Joined: Tue Nov 26, 2002 6:46 am
Location: Plainview, NY

Post by Another Peter »

The SWR is THE most important topic, but this approach, where you decide that the current value is above or below historical P/E ratioes looks to me as though you are trying to decide what will happen to your equity investments over the next short term period. This has proven to be very unsuccessful in the past. Can we change this to a different perspective? Will more frequent rebalancing when the PE ratio is above historical rates, and therefore more funds taken out of equities when the P/E is high be a dependable way to handle distributions?

Although you have five years in cash or near cash, you could still take this years distribution from equities if they increase in value.

Having said that, this is a innovative approach that should be used by all prior to FIRE, and adds to the importance of periodic rebalancing by asset class.
peteyperson
**** Heavy Hitter
Posts: 525
Joined: Tue Nov 26, 2002 6:46 am

Post by peteyperson »

Hi,

It is simply selling high and avoiding selling low by holding cash. It is not an attempt at forecasting, it just a realistically honest assessment of your net worth at that time (excluding out speculation in the pricing). Adjusting the valuation based on intrinsic valuation and not fairytale temporary ones good only the moment you price your shares/mutual funds is so we're all working from the same sheet of paper. It is not a physical share sale rebalancing issue. It is purely a paper process.

Seperating the adjustment from the w/d rate allows for everyone to reassess the value of their portfolio before looking at the SWR issue. Having then accounted for inflated values, the withdrawal rate is theoretically the same for all (if we all had the exact same asset allocation etc). The approach comes out the same as hocus in the end result, but it far easier IMHO to get your head around.

The actual rebalancing is not so much to go back to the original asset allocation, but to replenish any cash used during low valuation periods where your shares sales would be most costly. Deliberate rebalancing will have tax consequences for funds not sitting in tax-sheltered accounts. It is timing your sales to get the best value out of them. Better to sell at inflated values than lower than average P/E ratio, but I think you would still generally aim for the original asset allocation and not just sell a chunk because it is overvalued at the time (those investments sitting inside a tax-shelter can have much more flexibility in this issue).

It is a subtle revision of the hocus step-by-step approach aimed at simplicity and clarity. It is connected to the issue of SWR, but the SWR calculation is a different matter.

Petey



Another Peter wrote: The SWR is THE most important topic, but this approach, where you decide that the current value is above or below historical P/E ratioes looks to me as though you are trying to decide what will happen to your equity investments over the next short term period. This has proven to be very unsuccessful in the past. Can we change this to a different perspective? Will more frequent rebalancing when the PE ratio is above historical rates, and therefore more funds taken out of equities when the P/E is high be a dependable way to handle distributions?

Although you have five years in cash or near cash, you could still take this years distribution from equities if they increase in value.

Having said that, this is a innovative approach that should be used by all prior to FIRE, and adds to the importance of periodic rebalancing by asset class.
User avatar
ataloss
**** Heavy Hitter
Posts: 559
Joined: Mon Nov 25, 2002 3:00 am

Post by ataloss »

Hocus
I have not seen the work papers that William Bernstein used to
determine that the SWR for a high-stock portfolio at the top of the bull
market was 2 percent. But I trust that he did not pull a rabbit out of a hat
to come up with the number


Ah, but we really don't know where the number came from do we? I guess I am expecting something more than "I read this on page 234 of a book" given the talk about objectivity of SWR. Like:


hocus
SWR analysis, when it is true to what the data says, is different.
What the data says is not a matter of opinion. Data analysis involves
numbers. Data is hard, objective. You add up all the numbers that bear on the question being examined and you get a right answer to the question posed. That I find more helpful than someone's opinion of whether some investing idea is "good" or not.


you have formulated your idea of a unified theory of swr based on a number which is apparently an educated guess by Bernstein

I thought this exchange was interesting:

Gummy:
How exactly do you propose to take valuations into account? I don't see anything quantitative here, and I don't see how to get anything
quantitative.


Hocus:
There are a variety of reasonable ways to do the calculation. One
reasonable way is to follow the process used by Bernstein is coming to his
conclusion that the SWR in the year 2000 was 2 percent.


IOW, you have no idea how Bernstein got this number but you are willing to accept his "reasonable way" as fulfilling your "data based" swr objective.

Many experts are predicting lower future stock returns and posters from this board have made a strong case for this and a lower swr. People with actual experience with Monte Carlo simulation understand the limitations and sensitivity to the input.

Someone pointed out that risk adjustment had been mentioned on the tmf board. Petey is offended that I pointed out that intercst mentioned this to you > 1 year ago. I have no problem with Petey independently coming to intercst's conclusion but I question how much "progress" has been made since then. I see a lot of spinning of wheels. I see people on this board avoiding these threads (most being too polite I think)

hocus
I am saying that the tool should be set up so that it provides objective results. I am saying that to define what an SWR is so loosely that it can produce anything under the sun as a result is not a good idea because the tool becomes just one more way to engage in subjective evaluations of risk.


unfortunately the "tool" and the objective data don't exist
:wink:
Have fun.

Ataloss
Post Reply