Withdrawal Plan. McClung or Allocation Bands?

Have a question about your personal investments? No matter how simple or complex, you can ask it here.
Post Reply
Topic Author
HappyJack
Posts: 87
Joined: Thu Jan 10, 2019 7:45 am

Withdrawal Plan. McClung or Allocation Bands?

Post by HappyJack » Sat Oct 12, 2019 8:48 am

My original de-accumulation plan was to continually re-balance to keep my asset allocation within tolerance bands. I recently read Mr. Mcclung’s book. His idea is to sell only from bonds until the stock allocation is up 20% from the initial starting point. I would really like to know some peoples thoughts on these two withdrawal methods. I know that there are some threads on this, so if you would point me to those I would appreciate that also. Thank you.

RadAudit
Posts: 3584
Joined: Mon May 26, 2008 10:20 am
Location: Second star on the right and straight on 'til morning

Re: Withdrawal Plan. McClung or Allocation Bands?

Post by RadAudit » Sat Oct 12, 2019 10:27 am

HappyJack wrote:
Sat Oct 12, 2019 8:48 am
I know that there are some threads on this, so if you would point me to those I would appreciate that also.
Here's one. viewtopic.php?t=192105 Try the search window for others.
HappyJack wrote:
Sat Oct 12, 2019 8:48 am
His idea is to sell only from bonds until the stock allocation is up 20% from the initial starting point. I would really like to know some peoples thoughts on these two withdrawal methods.
The idea is, AFAIK, that you retire at a certain number and AA that you are reasonably certain will be sufficient to meet your needs for the duration. (If not, why retire?) Then, because stocks are more volatile than bonds, you withdraw from bonds and then rebalance - if stocks are up at least 20% plus inflation. The effect is to mitigate the probability of the adverse effects of sequence of return risk - i.e.: going broke. (Or, more simply, never sell stocks at a loss [below their initial value at retirement] in retirement in order to rebalance.*) I've got it in my IPS and had to use it once in retirement. Seems to work. I'm probably wrong on the explanation; but, I'm still not broke, either - yet.

*Edit: To eat is an equine of a different hue.
FI is the best revenge. LBYM. Invest the rest. Stay the course. - PS: The cavalry isn't coming, kids. You are on your own.

sport
Posts: 8443
Joined: Tue Feb 27, 2007 3:26 pm
Location: Cleveland, OH

Re: Withdrawal Plan. McClung or Allocation Bands?

Post by sport » Sat Oct 12, 2019 12:48 pm

If you let your stock allocation increase by 20%, you have a significantly higher level of risk. Consider what would occur if the stock market dropped 50% when your stock allocation had increased by 19%. I prefer to keep my stock allocation within 5% of my target.

retiredjg
Posts: 38199
Joined: Thu Jan 10, 2008 12:56 pm

Re: Withdrawal Plan. McClung or Allocation Bands?

Post by retiredjg » Sat Oct 12, 2019 12:51 pm

I would not let my stock allocation go up by 20% unless it started at something like 15% less than my comfort level.

NotWhoYouThink
Posts: 2713
Joined: Fri Dec 26, 2014 4:19 pm

Re: Withdrawal Plan. McClung or Allocation Bands?

Post by NotWhoYouThink » Sat Oct 12, 2019 1:19 pm

Not the stock allocation, the value of the equities in the portfolio.

Stocks going from $1M to $1.2M, not stocks going from 40% to 60% of the portfolio.

retiredjg
Posts: 38199
Joined: Thu Jan 10, 2008 12:56 pm

Re: Withdrawal Plan. McClung or Allocation Bands?

Post by retiredjg » Sat Oct 12, 2019 2:03 pm

Well, that is different, isn't it? I thought it sounded a bit strange...

RadAudit
Posts: 3584
Joined: Mon May 26, 2008 10:20 am
Location: Second star on the right and straight on 'til morning

Re: Withdrawal Plan. McClung or Allocation Bands?

Post by RadAudit » Sat Oct 12, 2019 3:06 pm

HappyJack wrote:
Sat Oct 12, 2019 8:48 am
His idea is to sell only from bonds until the stock allocation is up 20% from the initial starting point.
IIRC, not exactly. I believe the idea was to rebalance to the chosen AA if the value of stocks in the portfolio was higher than 20% from the starting point- not the stock allocation. Might help to run some numbers to see how it works.

Note: I'm lousy at math!! You'll have to run your own examples. But if you post your calculations - I'm sure someone will comment.
FI is the best revenge. LBYM. Invest the rest. Stay the course. - PS: The cavalry isn't coming, kids. You are on your own.

John Z
Posts: 386
Joined: Sun Nov 14, 2010 5:42 pm

Re: Withdrawal Plan. McClung or Allocation Bands?

Post by John Z » Sat Oct 12, 2019 7:34 pm

I've been using McClung's Prime Harvesting method since his book was published, with these parameters:

I went into retirement nearly 8 years ago with a slightly modified Coffeehouse Portfolio (60/40) that has 6 index stock funds and a total bond fund so it was easy to adapt my withdrawal strategy to the Prime Harvesting method. But rather than waiting for stock funds to go up 20% from the original amount and then convert to bonds (with the danger that stocks would increase but not reach 20% before crashing), I use a convert-to-bonds-at-a-10%-level as I wanted my allocation to shift to 40/60 in retirement and take the profits along the way. This has worked extremely well for me because my allocation now is 38/62 and I have taken annual withdrawals/RMDs along the way and have a larger balance than when I first retired (not due to my skill but the great market performance). The only stock fund I haven't withdrawn from or converted any amount to bonds is Total International Index since it has only grown about 5% since I retired. Following another one of his rules, I'll never buy stocks again.

User avatar
firebirdparts
Posts: 298
Joined: Thu Jun 13, 2019 4:21 pm

Re: Withdrawal Plan. McClung or Allocation Bands?

Post by firebirdparts » Sat Oct 12, 2019 9:12 pm

To me, it’s pretty obvious that you want to take some profits when stocks are soaring. Rebalancing will do that. You can define soaring and profits any way you want and back test whatever fancier method you want. The definitions used in this strategy are pretty plain vanilla. There isn’t much reason to criticize the concept.
A fool and your money are soon partners

Uncorrelated
Posts: 105
Joined: Sun Oct 13, 2019 3:16 pm

Re: Withdrawal Plan. McClung or Allocation Bands?

Post by Uncorrelated » Sun Oct 13, 2019 3:30 pm

There are plenty of reasons to criticize the concept, it's essentially market timing and tries very hard to hide that fact. Longinvest has written more on the topic: viewtopic.php?f=10&t=192105&p=2998394&h ... g#p2998394


My personal opinion is that if you believe in market timing, you should use your market timing indicators to create a path independent asset allocation. McClung's prime harvesting is no such strategy and as far as I'm aware of, no such retirement strategy has been proposed by anyone. The alternative is that you don't believe in market timing. If you don't believe in market timing, then you should choose either a constant asset allocation or a dynamic asset allocation that depends only on your current wealth and goal, because those are the only two possible strategies that are not market timing.

RadAudit
Posts: 3584
Joined: Mon May 26, 2008 10:20 am
Location: Second star on the right and straight on 'til morning

Re: Withdrawal Plan. McClung or Allocation Bands?

Post by RadAudit » Mon Oct 14, 2019 2:53 pm

Uncorrelated wrote:
Sun Oct 13, 2019 3:30 pm
My personal opinion is that if you believe in market timing, you should use your market timing indicators to create a path independent asset allocation. McClung's prime harvesting is no such strategy and as far as I'm aware of, no such retirement strategy has been proposed by anyone.
Uncorrelated wrote:
Sun Oct 13, 2019 3:30 pm
it's essentially market timing and tries very hard to hide that fact
Not to get in to a cat fight fight on this one; but, I'm having a little trouble in understanding this. I concur that I haven't seen too many strategies like prime harvesting in any place other than "Living Off Your Money." Could of missed it. But, I missed on how it's market timing in the classic sense. It essentially says to me - if you are in retirement and after a major drop in stock prices - don't start to rebalance until you have more money in stocks than what you had at the start. I'm guessing to do so would drop the total value of your portfolio below the number you decided was enough to last through retirement and probably would take longer to get back above what was considered enough when you started retirement, Don't know.
FI is the best revenge. LBYM. Invest the rest. Stay the course. - PS: The cavalry isn't coming, kids. You are on your own.

User avatar
vineviz
Posts: 5391
Joined: Tue May 15, 2018 1:55 pm

Re: Withdrawal Plan. McClung or Allocation Bands?

Post by vineviz » Mon Oct 14, 2019 4:04 pm

RadAudit wrote:
Mon Oct 14, 2019 2:53 pm
Uncorrelated wrote:
Sun Oct 13, 2019 3:30 pm
My personal opinion is that if you believe in market timing, you should use your market timing indicators to create a path independent asset allocation. McClung's prime harvesting is no such strategy and as far as I'm aware of, no such retirement strategy has been proposed by anyone.
Uncorrelated wrote:
Sun Oct 13, 2019 3:30 pm
it's essentially market timing and tries very hard to hide that fact
Not to get in to a cat fight fight on this one; but, I'm having a little trouble in understanding this. I concur that I haven't seen too many strategies like prime harvesting in any place other than "Living Off Your Money." Could of missed it. But, I missed on how it's market timing in the classic sense. It essentially says to me - if you are in retirement and after a major drop in stock prices - don't start to rebalance until you have more money in stocks than what you had at the start. I'm guessing to do so would drop the total value of your portfolio below the number you decided was enough to last through retirement and probably would take longer to get back above what was considered enough when you started retirement, Don't know.
It’s a strategy that changes asset allocation based entirely on the behavior of markets. For better or worse, that’s market timing.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch

dcabler
Posts: 1113
Joined: Wed Feb 19, 2014 11:30 am

Re: Withdrawal Plan. McClung or Allocation Bands?

Post by dcabler » Mon Oct 14, 2019 5:13 pm

vineviz wrote:
Mon Oct 14, 2019 4:04 pm
RadAudit wrote:
Mon Oct 14, 2019 2:53 pm
Uncorrelated wrote:
Sun Oct 13, 2019 3:30 pm
My personal opinion is that if you believe in market timing, you should use your market timing indicators to create a path independent asset allocation. McClung's prime harvesting is no such strategy and as far as I'm aware of, no such retirement strategy has been proposed by anyone.
Uncorrelated wrote:
Sun Oct 13, 2019 3:30 pm
it's essentially market timing and tries very hard to hide that fact
Not to get in to a cat fight fight on this one; but, I'm having a little trouble in understanding this. I concur that I haven't seen too many strategies like prime harvesting in any place other than "Living Off Your Money." Could of missed it. But, I missed on how it's market timing in the classic sense. It essentially says to me - if you are in retirement and after a major drop in stock prices - don't start to rebalance until you have more money in stocks than what you had at the start. I'm guessing to do so would drop the total value of your portfolio below the number you decided was enough to last through retirement and probably would take longer to get back above what was considered enough when you started retirement, Don't know.
It’s a strategy that changes asset allocation based entirely on the behavior of markets. For better or worse, that’s market timing.
I always thought market timing was based on a prediction of what markets are going to do, not what they've already done.

Investopdia uses this definition:
"What Is Market Timing?
Market timing is a type of investment or trading strategy. It is the act of moving in and out of a financial market or switching between asset classes based on predictive methods. These predictive tools include following technical indicators or economic data, to gauge how the market is going to move."

Uncorrelated
Posts: 105
Joined: Sun Oct 13, 2019 3:16 pm

Re: Withdrawal Plan. McClung or Allocation Bands?

Post by Uncorrelated » Mon Oct 14, 2019 5:36 pm

RadAudit wrote:
Mon Oct 14, 2019 2:53 pm
Uncorrelated wrote:
Sun Oct 13, 2019 3:30 pm
My personal opinion is that if you believe in market timing, you should use your market timing indicators to create a path independent asset allocation. McClung's prime harvesting is no such strategy and as far as I'm aware of, no such retirement strategy has been proposed by anyone.
Uncorrelated wrote:
Sun Oct 13, 2019 3:30 pm
it's essentially market timing and tries very hard to hide that fact
Not to get in to a cat fight fight on this one; but, I'm having a little trouble in understanding this. I concur that I haven't seen too many strategies like prime harvesting in any place other than "Living Off Your Money." Could of missed it. But, I missed on how it's market timing in the classic sense. It essentially says to me - if you are in retirement and after a major drop in stock prices - don't start to rebalance until you have more money in stocks than what you had at the start. I'm guessing to do so would drop the total value of your portfolio below the number you decided was enough to last through retirement and probably would take longer to get back above what was considered enough when you started retirement, Don't know.
Waiting to rebalance because stocks went down is an act of market timing. The root of the problem here is the path dependent property. Suppose that we are following a hypothetical withdrawal strategy: Start 50/50 stocks/bonds. Do not rebalance. Withdrawal from bonds first. If bonds are empty, withdrawal from stocks. This is the simplest strategy I can come up with that shows the problem.


Suppose that we have a retiree Bob. Bob was born in 1950 and retired in 2000 with $1M capital. He allocates $500k to stocks and $500k to bonds. It just so happens that in the first year of Bob's retirement, stocks do poorly and bonds to very well, after the first year he ends up with $400k in stocks and $600k in bonds.

There also is a different retiree Jane. Jane was also born in 1950 and retires in 2001 with $1M capital, using the exact same withdrawal strategy as Bob because she heard fro bob that this particular strategy was the best one. Jane allocates $500k to stocks and $500k to bonds


At this point, we have arrived at a paradox: there are two different retirees. They have the same age, the same risk appetite, the same remaining lifespan, and the same withdrawal strategy. Common sense should indicate that the optimal asset allocation for Bob is equal to the optimal asset allocation for Jane. But we can clearly see that Bob has an asset allocation of $400k/$600k and Jane has $500k/500k. What is the justification for Jane to take more risk than Bob? We say that the strategy is path dependent. There is a theory (the Principe of optimality) that states that path dependent strategies can not be optimal.

Intuitively it seems that all path dependent strategies depend on market timing in some way or another, but it is often far from clear why (even glidepaths and target date funds fail the test and are technically market timing). I don't know how to explain this, but it becomes quite obvious when you try to apply dynamic programming to asset allocation problems.


Just because it's market timing doesn't mean it won't work. But hiding the fact that it's market timing won't do any good. I have calculated optimal non-market timing strategies for various situations using dynamic programming techniques and will be creating a topic to discuss them some time in the future.

User avatar
vineviz
Posts: 5391
Joined: Tue May 15, 2018 1:55 pm

Re: Withdrawal Plan. McClung or Allocation Bands?

Post by vineviz » Mon Oct 14, 2019 5:37 pm

dcabler wrote:
Mon Oct 14, 2019 5:13 pm
I always thought market timing was based on a prediction of what markets are going to do, not what they've already done.
Relying on explicit market forecasts is one form (an especially strong form of market timing), but it's not the only form.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch

RadAudit
Posts: 3584
Joined: Mon May 26, 2008 10:20 am
Location: Second star on the right and straight on 'til morning

Re: Withdrawal Plan. McClung or Allocation Bands?

Post by RadAudit » Mon Oct 14, 2019 5:55 pm

Uncorrelated wrote:
Mon Oct 14, 2019 5:36 pm
I have calculated optimal non-market timing strategies for various situations using dynamic programming techniques and will be creating a topic to discuss them some time in the future.
Looking forward to it. Best of luck.
FI is the best revenge. LBYM. Invest the rest. Stay the course. - PS: The cavalry isn't coming, kids. You are on your own.

User avatar
samsoes
Posts: 1348
Joined: Tue Mar 05, 2013 9:12 am
Location: Northeast Rat Race

Re: Withdrawal Plan. McClung or Allocation Bands?

Post by samsoes » Mon Oct 14, 2019 6:00 pm

Sorry, I haven't read McClung's book...

When saying stocks up 20%, is that a relative 20%?

For example, in a 50/50 portfolio, stocks rising a relative 20% would mean stocks have reached 60% of the portfolio (50*1.2). Or does the 20% increase in stocks refer to an absolute 20%, meaning stocks have reached 70% of the portfolio (50+20)?

Thanks!
"Happiness Is Not My Companion" - Gen. Gouverneur K. Warren. | (Avatar is the statue of Gen. Warren atop Little Round Top @ Gettysburg National Military Park.)

Barsoom
Posts: 181
Joined: Thu Dec 06, 2018 9:40 am

Re: Withdrawal Plan. McClung or Allocation Bands?

Post by Barsoom » Mon Oct 14, 2019 7:01 pm

Uncorrelated wrote:
Mon Oct 14, 2019 5:36 pm
There also is a different retiree Jane. Jane was also born in 1950 and retires in 2001 with $1M capital, using the exact same withdrawal strategy as Bob because she heard fro bob that this particular strategy was the best one. Jane allocates $500k to stocks and $500k to bonds


At this point, we have arrived at a paradox: there are two different retirees. They have the same age, the same risk appetite, the same remaining lifespan, and the same withdrawal strategy. Common sense should indicate that the optimal asset allocation for Bob is equal to the optimal asset allocation for Jane. But we can clearly see that Bob has an asset allocation of $400k/$600k and Jane has $500k/500k. What is the justification for Jane to take more risk than Bob?
I've seen this hypothetical before, and I believe there is a small flaw in the logic. I can't say that the flaw is commensurate with the difference in portfolio, but we'll see.

Here's what I think the flaw is... "They have the same age, the same risk appetite, the same remaining lifespan, and the same withdrawal strategy," but not the same retirement length. Why is Jane starting at Bob's 50/50 allocation, given that she will have one less year of retirement than Bob? Bob started at 50/50, but he expected to fund a retirement that was one year longer than Jane's. If Jane has "the same risk appetite" as Bob and knows she's retiring into a down market from Bob, wouldn't Jane make a different starting allocation decision that factors in a shorter retirement period and a down market?

-B

NotWhoYouThink
Posts: 2713
Joined: Fri Dec 26, 2014 4:19 pm

Re: Withdrawal Plan. McClung or Allocation Bands?

Post by NotWhoYouThink » Mon Oct 14, 2019 7:04 pm

NotWhoYouThink wrote:
Sat Oct 12, 2019 1:19 pm
Not the stock allocation, the value of the equities in the portfolio.

Stocks going from $1M to $1.2M, not stocks going from 40% to 60% of the portfolio.
samsoes wrote:
Mon Oct 14, 2019 6:00 pm
Sorry, I haven't read McClung's book...

When saying stocks up 20%, is that a relative 20%?

For example, in a 50/50 portfolio, stocks rising a relative 20% would mean stocks have reached 60% of the portfolio (50*1.2). Or does the 20% increase in stocks refer to an absolute 20%, meaning stocks have reached 70% of the portfolio (50+20)?

Thanks!

randomguy
Posts: 8325
Joined: Wed Sep 17, 2014 9:00 am

Re: Withdrawal Plan. McClung or Allocation Bands?

Post by randomguy » Mon Oct 14, 2019 9:23 pm

Barsoom wrote:
Mon Oct 14, 2019 7:01 pm
Uncorrelated wrote:
Mon Oct 14, 2019 5:36 pm
There also is a different retiree Jane. Jane was also born in 1950 and retires in 2001 with $1M capital, using the exact same withdrawal strategy as Bob because she heard fro bob that this particular strategy was the best one. Jane allocates $500k to stocks and $500k to bonds


At this point, we have arrived at a paradox: there are two different retirees. They have the same age, the same risk appetite, the same remaining lifespan, and the same withdrawal strategy. Common sense should indicate that the optimal asset allocation for Bob is equal to the optimal asset allocation for Jane. But we can clearly see that Bob has an asset allocation of $400k/$600k and Jane has $500k/500k. What is the justification for Jane to take more risk than Bob?
I've seen this hypothetical before, and I believe there is a small flaw in the logic. I can't say that the flaw is commensurate with the difference in portfolio, but we'll see.

Here's what I think the flaw is... "They have the same age, the same risk appetite, the same remaining lifespan, and the same withdrawal strategy," but not the same retirement length. Why is Jane starting at Bob's 50/50 allocation, given that she will have one less year of retirement than Bob? Bob started at 50/50, but he expected to fund a retirement that was one year longer than Jane's. If Jane has "the same risk appetite" as Bob and knows she's retiring into a down market from Bob, wouldn't Jane make a different starting allocation decision that factors in a shorter retirement period and a down market?

-B
Do you think the differences between a 29 year retirement and a 30 year one is enough to change ones AA that much? Now overweighting stocks because they are down 20% might be a good move, but it is pure market timing.

My question is always what do you do when you retire in bad times. Take the 2000 retiree who was 50/50 S&P 500 and bonds. From 2000-2011, the market wasn't up 20%. Is it really ok to hit 2012 with like 14x in stocks and 2x in bonds in nominal dollars? With any of these schemes, you should plot out how they worked in 1929, 1966, 1973, and 2000 to get a feel of how comfortable you are with them. Pretty much anything looks great during average markets. It is only when you get a horrid 10+ years do you get stuck having to make really tough decisions.

It should be pointed out that in a normal fixed AA, you basically never sell stocks when they are down. You sell bonds and buy stocks. It takes a really long down market before you are selling stocks for a loss versus either selling bonds or stocks you bought at lower prices. If you do better by not rebalancing or not is very situation dependent. Sometimes not rebalancing helps. Other times it hurts. It tends to make a very small difference.

Topic Author
HappyJack
Posts: 87
Joined: Thu Jan 10, 2019 7:45 am

Re: Withdrawal Plan. McClung or Allocation Bands?

Post by HappyJack » Mon Oct 14, 2019 10:30 pm

OP here
My plan has been to rebalance often in retirement using tolerance bands in each asset class (fund or etf). I read a Vanguard paper by Daryanani that seemed to make sense. Is this a good approach in retirement/de-accumulation if coupled with VPW withdrawal system. Thoughts?

Topic Author
HappyJack
Posts: 87
Joined: Thu Jan 10, 2019 7:45 am

Re: Withdrawal Plan. McClung or Allocation Bands?

Post by HappyJack » Mon Oct 14, 2019 10:31 pm

OP here
My plan has been to rebalance often in retirement using tolerance bands in each asset class (fund or etf). I read a Vanguard paper by Daryanani that seemed to make sense. Is this a good approach in retirement/de-accumulation if coupled with VPW withdrawal system. Thoughts?

marcopolo
Posts: 2399
Joined: Sat Dec 03, 2016 10:22 am

Re: Withdrawal Plan. McClung or Allocation Bands?

Post by marcopolo » Tue Oct 15, 2019 12:10 am

NotWhoYouThink wrote:
Mon Oct 14, 2019 7:04 pm
NotWhoYouThink wrote:
Sat Oct 12, 2019 1:19 pm
Not the stock allocation, the value of the equities in the portfolio.

Stocks going from $1M to $1.2M, not stocks going from 40% to 60% of the portfolio.
samsoes wrote:
Mon Oct 14, 2019 6:00 pm
Sorry, I haven't read McClung's book...

When saying stocks up 20%, is that a relative 20%?

For example, in a 50/50 portfolio, stocks rising a relative 20% would mean stocks have reached 60% of the portfolio (50*1.2). Or does the 20% increase in stocks refer to an absolute 20%, meaning stocks have reached 70% of the portfolio (50+20)?

Thanks!

I would add that McClung's 20% increase in value of your equities to trigger re-balancing is supposed to be inflation adjusted.

So, if you start with $1M in equities, you would re-balance only when your equities become $1.2M after adjusting for inflation. So, if you had 20% inflation in the intervening years, the trigger point would more like ~$1.44M in equities.
Once in a while you get shown the light, in the strangest of places if you look at it right.

Barsoom
Posts: 181
Joined: Thu Dec 06, 2018 9:40 am

Re: Withdrawal Plan. McClung or Allocation Bands?

Post by Barsoom » Tue Oct 15, 2019 12:50 am

randomguy wrote:
Mon Oct 14, 2019 9:23 pm
Do you think the differences between a 29 year retirement and a 30 year one is enough to change ones AA that much? Now overweighting stocks because they are down 20% might be a good move, but it is pure market timing.
No, I don't. But if we're making hypotheticals and calling them equal, I'm calling out where they are not equal.
randomguy wrote:
Mon Oct 14, 2019 9:23 pm
My question is always what do you do when you retire in bad times.
That's the situation in the hypothetical. Jane retires into a down market. Would she blindly have the same AA as Bob, who otherwise has the same demographic? Maybe, if she thinks the market has bottomed and will rise again. If she thinks it's still falling, then why invest in a falling market? Shouldn't she overweight bonds/cash before entering the market for the first time?

In fact, is it "market timing" to choose when to first enter the market, or is it prudent to recognize a down market and wait for signs of growth before investing those first dollars?
randomguy wrote:
Mon Oct 14, 2019 9:23 pm
It should be pointed out that in a normal fixed AA, you basically never sell stocks when they are down. You sell bonds and buy stocks. It takes a really long down market before you are selling stocks for a loss versus either selling bonds or stocks you bought at lower prices. If you do better by not rebalancing or not is very situation dependent. Sometimes not rebalancing helps. Other times it hurts. It tends to make a very small difference.
That's the McClung strategy, except that he never buys stocks again. In fact, as a retiree does one still buy stocks in a decumulation phase?

-B

randomguy
Posts: 8325
Joined: Wed Sep 17, 2014 9:00 am

Re: Withdrawal Plan. McClung or Allocation Bands?

Post by randomguy » Tue Oct 15, 2019 1:07 am

marcopolo wrote:
Tue Oct 15, 2019 12:10 am

I would add that McClung's 20% increase in value of your equities to trigger re-balancing is supposed to be inflation adjusted.

So, if you start with $1M in equities, you would re-balance only when your equities become $1.2M after adjusting for inflation. So, if you had 20% inflation in the intervening years, the trigger point would more like ~$1.44M in equities.
So the 2000 retiree would need to go from jan 2000 to October 2013 by spending bonds? Granted that is one of the worst periods every but that might be pretty tough on the heart. :)

There is no magic in any of these schemes. Not selling your stocks when they are down does not avoid sequence of return risk. Even if you don't sell a share, ending year 13 with 1.2 million instead of the 2 million you expect in average cases still exposes you to SOR. The lost opportunity cost still gets you.

Depending on the scheme, you might end up with slightly better returns (you have more stocks when returns are good) or bad (you have less stocks when returns are good). It tends not to be a big difference except in extreme conditions where you either rebalance all your money away (imagine 15 years of a falling market) or you catch a bounce (imagine stocks drop 75% and then rebound by 75%. The person who rebalanced has more stocks to participate in the gains).

They are far from horrible schemes but they just don't do much over just holding 30/70 to 60/40. There seems to be two types of people
a) those who like simplicity. By not having a lot of choices,they are happy with how things go.
b) those who like complexity. Choices make them happy since they feel like they are doing something.

Both work ok. You just need to pick which one satisfies your mental needs..

randomguy
Posts: 8325
Joined: Wed Sep 17, 2014 9:00 am

Re: Withdrawal Plan. McClung or Allocation Bands?

Post by randomguy » Tue Oct 15, 2019 1:13 am

Barsoom wrote:
Tue Oct 15, 2019 12:50 am
randomguy wrote:
Mon Oct 14, 2019 9:23 pm
My question is always what do you do when you retire in bad times.
That's the situation in the hypothetical. Jane retires into a down market. Would she blindly have the same AA as Bob, who otherwise has the same demographic? Maybe, if she thinks the market has bottomed and will rise again. If she thinks it's still falling, then why invest in a falling market? Shouldn't she overweight bonds/cash before entering the market for the first time?

In fact, is it "market timing" to choose when to first enter the market, or is it prudent to recognize a down market and wait for signs of growth before investing those first dollars?
A year when stocks are down 20% isn't bad times. That is a normal occurrence that you should expect to happen on a regular basis. Bad times are when you retire and stocks are down for 10-15 years. These schemes are try to make the first case easier to deal with but they do nothing about the 2nd case.

Sure if you time the market right, waiting for growth to start is prudent. Problem is when you don't buy back in. Between 2012-2016, this board was filled with posts from people who sold in 2008-9 and were asking if it was ok to buy back in now at our elevated valuations.

Uncorrelated
Posts: 105
Joined: Sun Oct 13, 2019 3:16 pm

Re: Withdrawal Plan. McClung or Allocation Bands?

Post by Uncorrelated » Tue Oct 15, 2019 3:41 am

Barsoom wrote:
Mon Oct 14, 2019 7:01 pm
Uncorrelated wrote:
Mon Oct 14, 2019 5:36 pm
There also is a different retiree Jane. Jane was also born in 1950 and retires in 2001 with $1M capital, using the exact same withdrawal strategy as Bob because she heard fro bob that this particular strategy was the best one. Jane allocates $500k to stocks and $500k to bonds


At this point, we have arrived at a paradox: there are two different retirees. They have the same age, the same risk appetite, the same remaining lifespan, and the same withdrawal strategy. Common sense should indicate that the optimal asset allocation for Bob is equal to the optimal asset allocation for Jane. But we can clearly see that Bob has an asset allocation of $400k/$600k and Jane has $500k/500k. What is the justification for Jane to take more risk than Bob?
Here's what I think the flaw is... "They have the same age, the same risk appetite, the same remaining lifespan, and the same withdrawal strategy," but not the same retirement length. Why is Jane starting at Bob's 50/50 allocation, given that she will have one less year of retirement than Bob?
That is indeed a flaw in the example. Let's say Bob targets a 31 year retirement (starting in 2000) and jane targets a 30 year retirement (starting in 2001). If 50/50 is the optimal allocation for Bob, why is that also the optimal allocation for Jane? Maybe the optimal allocation for Jane is 55/45.

But that results in another paradox: Bob starts with an asset allocation of 50/50. After the first year, market movements have caused his portfolio to change to 40/60 and his remaining retirement duration is 30 year. But we just said that the optimal allocation for a 30-year retirement is 55/45! That can't be right.


Path independence says: if we have a 31 year retirement duration, then the optimal solution for the last 30 years is equal to the optimal solution for a 30 year retirement.
Jane retires into a down market.
This is one of the thoughts that often results in market timing. If market timing is impossible, then it doesn't matter if we retired during a bull market or a bear market. The future outlook is the same in both cases.
In fact, is it "market timing" to choose when to first enter the market, or is it prudent to recognize a down market and wait for signs of growth before investing those first dollars?
Definitely market timing. See argument above.

Uncorrelated
Posts: 105
Joined: Sun Oct 13, 2019 3:16 pm

Re: Withdrawal Plan. McClung or Allocation Bands?

Post by Uncorrelated » Tue Oct 15, 2019 3:44 am

randomguy wrote:
Tue Oct 15, 2019 1:07 am
There is no magic in any of these schemes. Not selling your stocks when they are down does not avoid sequence of return risk. Even if you don't sell a share, ending year 13 with 1.2 million instead of the 2 million you expect in average cases still exposes you to SOR. The lost opportunity cost still gets you.
Ding ding. We have a winner.

User avatar
Raybo
Posts: 1846
Joined: Tue Feb 20, 2007 11:02 am
Location: San Francisco
Contact:

Re: Withdrawal Plan. McClung or Allocation Bands?

Post by Raybo » Tue Oct 15, 2019 8:14 am

Bob has $1M in 2000 allocated at 50/50 and after a year he is at 40/60 but still has $1M. In 2001, Jane has $1M and an AA of 50/50. What was Jane's AA in 2000?

500K / .8 = $625K in stocks

500K / 1.2 = $416K in bonds.

Jane had $1.41M in 2000 and an AA of 60/40. If Jane sets her AA at 50/50 when she retires, then either her AA was to high at 60/40 or she, too, is market timing by not rebalancing.

While these 2 people might have the same amount of money at the start of 2001, I don't see how you can say they are equivalent.
No matter how long the hill, if you keep pedaling you'll eventually get up to the top.

sixtyforty
Posts: 362
Joined: Tue Nov 25, 2014 12:22 pm
Location: USA

Re: Withdrawal Plan. McClung or Allocation Bands?

Post by sixtyforty » Tue Oct 15, 2019 8:47 am

I read Mclungs book and IMO it feels a bit like data mining, even though he states in the beginning of the book he tries to avoid it.
"Simplicity is the ultimate sophistication" - Leonardo Da Vinci

Uncorrelated
Posts: 105
Joined: Sun Oct 13, 2019 3:16 pm

Re: Withdrawal Plan. McClung or Allocation Bands?

Post by Uncorrelated » Tue Oct 15, 2019 8:49 am

Raybo wrote:
Tue Oct 15, 2019 8:14 am
Bob has $1M in 2000 allocated at 50/50 and after a year he is at 40/60 but still has $1M. In 2001, Jane has $1M and an AA of 50/50. What was Jane's AA in 2000?
That information isn't relevant to the question. Maybe she was still working in 2000.
While these 2 people might have the same amount of money at the start of 2001, I don't see how you can say they are equivalent.
This is the result of behavioral bias that is known as sunk cost fallacy. Money is fungible. It doesn't matter if you have 100% your money in cash, in stocks or in peanut butter futures. The answer to the question "what is the optimal asset allocation?" depends only on future goals and expectations, not on your previous asset allocation.

User avatar
Raybo
Posts: 1846
Joined: Tue Feb 20, 2007 11:02 am
Location: San Francisco
Contact:

Re: Withdrawal Plan. McClung or Allocation Bands?

Post by Raybo » Tue Oct 15, 2019 9:41 am

Uncorrelated wrote:
Tue Oct 15, 2019 8:49 am
Raybo wrote:
Tue Oct 15, 2019 8:14 am
Bob has $1M in 2000 allocated at 50/50 and after a year he is at 40/60 but still has $1M. In 2001, Jane has $1M and an AA of 50/50. What was Jane's AA in 2000?
That information isn't relevant to the question. Maybe she was still working in 2000.
While these 2 people might have the same amount of money at the start of 2001, I don't see how you can say they are equivalent.
This is the result of behavioral bias that is known as sunk cost fallacy. Money is fungible. It doesn't matter if you have 100% your money in cash, in stocks or in peanut butter futures. The answer to the question "what is the optimal asset allocation?" depends only on future goals and expectations, not on your previous asset allocation.
I disagree.

Jane has lost 4% of her asset pile in 2000. I assume that Bob spent 4% of his asset pile in 2000, yet, according to your story, he still has $1M. They now both have the same amount of money.

Jane had a more aggressive AA in 2000 and she still has a more aggressive AA in 2001. Clearly, their desires to take risk are different. Maybe Jane wants to leave a legacy or take expensive vacations. Maybe Bob's parents lived through the depression and taught him extreme frugality. Who knows?

Since these two people have different investment strategies, identifying which one is "optimal" isn't possible.

What you are saying is that going forward 30 years, one of these two people will have a better outcome than the other. This is obvious by the fact that they have different AAs but the same amount of money. Will either run out of money? Who knows. Maybe peanut butter futures will turn out to be the best investment!
No matter how long the hill, if you keep pedaling you'll eventually get up to the top.

User avatar
vineviz
Posts: 5391
Joined: Tue May 15, 2018 1:55 pm

Re: Withdrawal Plan. McClung or Allocation Bands?

Post by vineviz » Tue Oct 15, 2019 9:51 am

Raybo wrote:
Tue Oct 15, 2019 9:41 am
Jane had a more aggressive AA in 2000 and she still has a more aggressive AA in 2001. Clearly, their desires to take risk are different. Maybe Jane wants to leave a legacy or take expensive vacations. Maybe Bob's parents lived through the depression and taught him extreme frugality. Who knows?
Remember that it was specified that Bob and Jane have the same level of risk aversion. Given that, if they have the same wealth and the same investment goal then they should have the the same asset allocation if they are both rational.

Any process that gives two identical people, who are in the the same current financial position, different asset allocations is path dependent and therefore flawed.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch

User avatar
Raybo
Posts: 1846
Joined: Tue Feb 20, 2007 11:02 am
Location: San Francisco
Contact:

Re: Withdrawal Plan. McClung or Allocation Bands?

Post by Raybo » Tue Oct 15, 2019 10:04 am

vineviz wrote:
Tue Oct 15, 2019 9:51 am
Raybo wrote:
Tue Oct 15, 2019 9:41 am
Jane had a more aggressive AA in 2000 and she still has a more aggressive AA in 2001. Clearly, their desires to take risk are different. Maybe Jane wants to leave a legacy or take expensive vacations. Maybe Bob's parents lived through the depression and taught him extreme frugality. Who knows?
Remember that it was specified that Bob and Jane have the same level of risk aversion. Given that, if they have the same wealth and the same investment goal then they should have the the same asset allocation if they are both rational.

Any process that gives two identical people, who are in the the same current financial position, different asset allocations is path dependent and therefore flawed.
The purpose of the hypothetical is to try to show something not shown by the example. How is it possible for two people to have the same level of risk aversion and, yet, have 2 different AA a priori? So, is one of them irrational or is the example flawed?
No matter how long the hill, if you keep pedaling you'll eventually get up to the top.

User avatar
vineviz
Posts: 5391
Joined: Tue May 15, 2018 1:55 pm

Re: Withdrawal Plan. McClung or Allocation Bands?

Post by vineviz » Tue Oct 15, 2019 10:20 am

Raybo wrote:
Tue Oct 15, 2019 10:04 am
vineviz wrote:
Tue Oct 15, 2019 9:51 am
Raybo wrote:
Tue Oct 15, 2019 9:41 am
Jane had a more aggressive AA in 2000 and she still has a more aggressive AA in 2001. Clearly, their desires to take risk are different. Maybe Jane wants to leave a legacy or take expensive vacations. Maybe Bob's parents lived through the depression and taught him extreme frugality. Who knows?
Remember that it was specified that Bob and Jane have the same level of risk aversion. Given that, if they have the same wealth and the same investment goal then they should have the the same asset allocation if they are both rational.

Any process that gives two identical people, who are in the the same current financial position, different asset allocations is path dependent and therefore flawed.
The purpose of the hypothetical is to try to show something not shown by the example. How is it possible for two people to have the same level of risk aversion and, yet, have 2 different AA a priori? So, is one of them irrational or is the example flawed?
If they are both setting their asset allocation based on McClung's method, they are arguably both being irrational.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch

RadAudit
Posts: 3584
Joined: Mon May 26, 2008 10:20 am
Location: Second star on the right and straight on 'til morning

Re: Withdrawal Plan. McClung or Allocation Bands?

Post by RadAudit » Tue Oct 15, 2019 10:55 am

randomguy wrote:
Mon Oct 14, 2019 9:23 pm

My question is always what do you do when you retire in bad times. Take the 2000 retiree who was 50/50 S&P 500 and bonds. From 2000-2011, the market wasn't up 20%. Is it really ok to hit 2012 with like 14x in stocks and 2x in bonds in nominal dollars? With any of these schemes, you should plot out how they worked in 1929, 1966, 1973, and 2000 to get a feel of how comfortable you are with them. Pretty much anything looks great during average markets. It is only when you get a horrid 10+ years do you get stuck having to make really tough decisions.

It should be pointed out that in a normal fixed AA, you basically never sell stocks when they are down. You sell bonds and buy stocks. It takes a really long down market before you are selling stocks for a loss versus either selling bonds or stocks you bought at lower prices. If you do better by not rebalancing or not is very situation dependent. Sometimes not rebalancing helps. Other times it hurts. It tends to make a very small difference.
Excellent point. I did some rough back of the envelope calculations. I don't think I'd feel too good with a 10 year run of poor luck in retirement while waiting for blind adherence to a theory to pull me out that situation. Unless I had a lot of bonds to start, a low SWR and a relatively short remaining life expectancy. Then maybe I could tough it out.
FI is the best revenge. LBYM. Invest the rest. Stay the course. - PS: The cavalry isn't coming, kids. You are on your own.

Uncorrelated
Posts: 105
Joined: Sun Oct 13, 2019 3:16 pm

Re: Withdrawal Plan. McClung or Allocation Bands?

Post by Uncorrelated » Tue Oct 15, 2019 11:03 am

Raybo wrote:
Tue Oct 15, 2019 10:04 am
vineviz wrote:
Tue Oct 15, 2019 9:51 am
Raybo wrote:
Tue Oct 15, 2019 9:41 am
Jane had a more aggressive AA in 2000 and she still has a more aggressive AA in 2001. Clearly, their desires to take risk are different. Maybe Jane wants to leave a legacy or take expensive vacations. Maybe Bob's parents lived through the depression and taught him extreme frugality. Who knows?
Remember that it was specified that Bob and Jane have the same level of risk aversion. Given that, if they have the same wealth and the same investment goal then they should have the the same asset allocation if they are both rational.

Any process that gives two identical people, who are in the the same current financial position, different asset allocations is path dependent and therefore flawed.
The purpose of the hypothetical is to try to show something not shown by the example. How is it possible for two people to have the same level of risk aversion and, yet, have 2 different AA a priori? So, is one of them irrational or is the example flawed?
The purpose of the hypothetical is to show that the strategy is path independent. A path independent strategy is not optimal (theory: the principle of optimality).

Maybe Jane was still working in 2000. Maybe Jane received a large inheritance in the end of 2000. Maybe Bob retired on 1 jan 2000 and Jane on 2 jan 2000 and the mercy of the markets resulted in widely different asset allocations in 2001. There are a million ways we can twist the example to make it more equal, but that doesn't change the conclusion that the strategy is path dependent.


The responses given by vineviz are on point.

marcopolo
Posts: 2399
Joined: Sat Dec 03, 2016 10:22 am

Re: Withdrawal Plan. McClung or Allocation Bands?

Post by marcopolo » Tue Oct 15, 2019 12:42 pm

randomguy wrote:
Tue Oct 15, 2019 1:07 am
marcopolo wrote:
Tue Oct 15, 2019 12:10 am

I would add that McClung's 20% increase in value of your equities to trigger re-balancing is supposed to be inflation adjusted.

So, if you start with $1M in equities, you would re-balance only when your equities become $1.2M after adjusting for inflation. So, if you had 20% inflation in the intervening years, the trigger point would more like ~$1.44M in equities.
So the 2000 retiree would need to go from jan 2000 to October 2013 by spending bonds? Granted that is one of the worst periods every but that might be pretty tough on the heart. :)

There is no magic in any of these schemes. Not selling your stocks when they are down does not avoid sequence of return risk. Even if you don't sell a share, ending year 13 with 1.2 million instead of the 2 million you expect in average cases still exposes you to SOR. The lost opportunity cost still gets you.

Depending on the scheme, you might end up with slightly better returns (you have more stocks when returns are good) or bad (you have less stocks when returns are good). It tends not to be a big difference except in extreme conditions where you either rebalance all your money away (imagine 15 years of a falling market) or you catch a bounce (imagine stocks drop 75% and then rebound by 75%. The person who rebalanced has more stocks to participate in the gains).

They are far from horrible schemes but they just don't do much over just holding 30/70 to 60/40. There seems to be two types of people
a) those who like simplicity. By not having a lot of choices,they are happy with how things go.
b) those who like complexity. Choices make them happy since they feel like they are doing something.

Both work ok. You just need to pick which one satisfies your mental needs..
I completely agree with you. The various withdrawal strategies largely just trade off one type of risk for another. There is no magic.

I was just pointing out what the actual proposal put forth by McClung was, as there seemed to be some misunderstanding earlier in the thread. I was not advocating for it.
Once in a while you get shown the light, in the strangest of places if you look at it right.

User avatar
2pedals
Posts: 913
Joined: Wed Dec 31, 2014 12:31 pm

Re: Withdrawal Plan. McClung or Allocation Bands?

Post by 2pedals » Tue Oct 15, 2019 1:50 pm

FYI
McClung book has two different types of "prime harvesting" after the stock value has reached target value of stock.


Prime Harvesting
rebalance stock-bond ratio by selling 20% of stocks to buy additional bonds

Alternate-Prime Harvesting
sell enough stocks to bring bonds back to initial target percentage


He recommends prime harvesting and to use target value as 120% times initial total stock value at retirement adjusted for inflation.

User avatar
2pedals
Posts: 913
Joined: Wed Dec 31, 2014 12:31 pm

Re: Withdrawal Plan. McClung or Allocation Bands?

Post by 2pedals » Tue Oct 15, 2019 9:08 pm

vineviz wrote:
Mon Oct 14, 2019 4:04 pm
RadAudit wrote:
Mon Oct 14, 2019 2:53 pm
Uncorrelated wrote:
Sun Oct 13, 2019 3:30 pm
My personal opinion is that if you believe in market timing, you should use your market timing indicators to create a path independent asset allocation. McClung's prime harvesting is no such strategy and as far as I'm aware of, no such retirement strategy has been proposed by anyone.
Uncorrelated wrote:
Sun Oct 13, 2019 3:30 pm
it's essentially market timing and tries very hard to hide that fact
Not to get in to a cat fight fight on this one; but, I'm having a little trouble in understanding this. I concur that I haven't seen too many strategies like prime harvesting in any place other than "Living Off Your Money." Could of missed it. But, I missed on how it's market timing in the classic sense. It essentially says to me - if you are in retirement and after a major drop in stock prices - don't start to rebalance until you have more money in stocks than what you had at the start. I'm guessing to do so would drop the total value of your portfolio below the number you decided was enough to last through retirement and probably would take longer to get back above what was considered enough when you started retirement, Don't know.
It’s a strategy that changes asset allocation based entirely on the behavior of markets. For better or worse, that’s market timing.
The "prime harvesting" withdrawal method does not directly change asset allocation. The AA is allowed to float based on an initial stock amount and market performance. If the stock amount exceeds a predetermined amount adjusted for inflation stocks are sold to get back to the original stock amount plus inflation.

When assuming a given amount, for fixed AA allocation when the market goes up you are selling stocks and when the market goes down are your buying stocks which is based on market behavior. When using prime harvesting you only selling stocks based on good stock market behavior.

Usually, in retirement your assets are already built and the primary goal is to sell stocks when spending your money down and many people would like a systematic method to do this. Prime harvesting is one way to do this.

It is okay, you can call it market timing but that perspective disregards the importance of the primary goals.

Uncorrelated
Posts: 105
Joined: Sun Oct 13, 2019 3:16 pm

Re: Withdrawal Plan. McClung or Allocation Bands?

Post by Uncorrelated » Wed Oct 16, 2019 2:52 am

2pedals wrote:
Tue Oct 15, 2019 9:08 pm
The "prime harvesting" withdrawal method does not directly change asset allocation. The AA is allowed to float based on an initial stock amount and market performance.
If at some point you have an asset allocation of 60/40 and the next point you have an asset allocation of 65/35 because of market movements, you are in fact changing your asset allocation. Even if you didn't take conscious action to make that happen.


It has been mathematically proven that if you cannot predict the market, then a constant allocation that is frequent rebalanced is the best possible asset allocation. If you ever change your asset allocation, then either your risk appetite has changed or you are market timing.


It is okay, you can call it market timing but that perspective disregards the importance of the primary goals.
The primary goal is market timing in an effort to make your retirement money last longer. Not saying that that necessary will not work, but that's just what it is.

From a theoretical perspective, if you can time the market, you should be timing the market by continuously changing your asset allocation to the optimal one. Not by floating your asset allocation at the mercy of the markets and making arbitrary breakpoints where the asset allocation is changed in a non-continuous way.

User avatar
2pedals
Posts: 913
Joined: Wed Dec 31, 2014 12:31 pm

Re: Withdrawal Plan. McClung or Allocation Bands?

Post by 2pedals » Wed Oct 16, 2019 10:21 am

Uncorrelated wrote:
Wed Oct 16, 2019 2:52 am
2pedals wrote:
Tue Oct 15, 2019 9:08 pm
The "prime harvesting" withdrawal method does not directly change asset allocation. The AA is allowed to float based on an initial stock amount and market performance.
If at some point you have an asset allocation of 60/40 and the next point you have an asset allocation of 65/35 because of market movements, you are in fact changing your asset allocation. Even if you didn't take conscious action to make that happen.


It has been mathematically proven that if you cannot predict the market, then a constant allocation that is frequent rebalanced is the best possible asset allocation. If you ever change your asset allocation, then either your risk appetite has changed or you are market timing.


It is okay, you can call it market timing but that perspective disregards the importance of the primary goals.
The primary goal is market timing in an effort to make your retirement money last longer. Not saying that that necessary will not work, but that's just what it is.

From a theoretical perspective, if you can time the market, you should be timing the market by continuously changing your asset allocation to the optimal one. Not by floating your asset allocation at the mercy of the markets and making arbitrary breakpoints where the asset allocation is changed in a non-continuous way.
Anybody using the opposite approach of market timing, just buying and holding stocks has the expectation at some point the stock market will go up. Prime harvesting only sells when this happens. Prime harvesting does not try to predict when the best time to be in or out of the market. If this is market timing, this is an extremely mild form of market timing.

One can argue that holding a fixed allocation when widthdrawing funds (as in retirement) is a form of market timing as well. You are buying and selling stocks based on market performance. You are not just buying and holding stocks. If stocks are going up many times you may need to sell to rebalance and you will need to sell to take withdrawals, (not hold). If the stock market goes down sometimes you may need to buy additional stocks that could increase risks during sharp market downturns. Many people like the attribute of using a fixed allocation since you are typically buying when stocks are going down and selling when stocks are going up. This is a very mild form of market timing as well.

randomguy
Posts: 8325
Joined: Wed Sep 17, 2014 9:00 am

Re: Withdrawal Plan. McClung or Allocation Bands?

Post by randomguy » Wed Oct 16, 2019 10:50 am

2pedals wrote:
Tue Oct 15, 2019 9:08 pm
It is okay, you can call it market timing but that perspective disregards the importance of the primary goals.
Ignore if you are market timing or not. How well does it do at meeting those goals compared to a fixed AA in 1929, 1966,1973, and 2000? Was there a noticeable difference? How did you feel when from 1966-1984 (or 1929-1943 or 2000-2012), you were selling bonds every year and your allocation when from say 40/60 to something like 100/0?

Now I am not saying fixed AA are perfect. Over time, you needs can change. A 85 year old with 50x in expenses (you had 20 years of decent returns) can change from the 40/60 they have been holding. They could put 15 years in bonds and shove the rest in stocks. They could go all bonds.

Uncorrelated
Posts: 105
Joined: Sun Oct 13, 2019 3:16 pm

Re: Withdrawal Plan. McClung or Allocation Bands?

Post by Uncorrelated » Wed Oct 16, 2019 1:10 pm

2pedals wrote:
Wed Oct 16, 2019 10:21 am
Anybody using the opposite approach of market timing, just buying and holding stocks has the expectation at some point the stock market will go up. Prime harvesting only sells when this happens. Prime harvesting does not try to predict when the best time to be in or out of the market. If this is market timing, this is an extremely mild form of market timing.
Prime harvesting says you shouldn't sell your stocks if they are up 19%, but you should if they are up 20%. Clearly, prime harvesting is saying that you should be in the market when stocks are up 19%, but not when they are up 20%.

Whether stocks are up compared to your arbitrary retirement day is not a good predictor of future stock returns.

A constant asset allocation doesn't predict, it takes the same risks every day.
One can argue that holding a fixed allocation when widthdrawing funds (as in retirement) is a form of market timing as well. You are buying and selling stocks based on market performance. You are not just buying and holding stocks. If stocks are going up many times you may need to sell to rebalance and you will need to sell to take withdrawals, (not hold). If the stock market goes down sometimes you may need to buy additional stocks that could increase risks during sharp market downturns. Many people like the attribute of using a fixed allocation since you are typically buying when stocks are going down and selling when stocks are going up. This is a very mild form of market timing as well.
See this paper, theory 1 on page 119: An Introduction to Computational Finance Without Agonizing Pain

They model the market with brownian motion with constant drift. Theory 1 states: suppose we have an asset allocation of 2 assets where the proportion between assets is not constant. Then It is possible to choose a linear asset allocation which has the same expected value but lower standard deviation.

In theory 2, they extend the problem to a jump diffusion model.

Surely, a strategy that is optimal on a model market that is unpredictable by definition, can not be a market timing strategy.

User avatar
2pedals
Posts: 913
Joined: Wed Dec 31, 2014 12:31 pm

Re: Withdrawal Plan. McClung or Allocation Bands?

Post by 2pedals » Wed Oct 16, 2019 11:41 pm

Uncorrelated wrote:
Wed Oct 16, 2019 1:10 pm
A constant allocation is an active strategy strategy because you are buy and selling based on market conditions. A buy and hold strategy while allowing for allocation drift is a passive strategy. A buy and hold strategy is not market timing. Market timing is when someone attempts to predict future market price movements.

Prime harvesting doesn't directly try to predict future market price movements. It is just a rule that says to sell some stocks if they go up. The only prediction is that if stocks go up above the target amount it is probably a good time to sell, yes it fine to take some your bets off the table. Furthermore the variable withdrawal methods recommended by McClung will tell you that is probably okay to spend more money than before.

A constant allocation strategy is based an implicit bet against diverging and poor stock asset performance for the duration of your life or assets. If while in the duration of your retirement stocks under perform relative to bonds you will be buying stocks while at the same withdrawing money for retirement. A constant allocation strategy if done during a diverging duration will perform worse than a buy and hold strategy.

In retirement I have a bucket of stocks (risky assets), bucket of bonds (not so risky assets) and income that should be more than enough to cover my retirement expenses. My initial asset allocation was set based on how much assets I feel comfortable placing in the risk bucket for the duration of my retirement. Risking additional stock in a possible diverging stock asset performance is a risk I believe I do not need to take in retirement. I like to sleep well at night. Having said that am I going to strictly follow McClung's recommendation? Nope. I will take risk money off the table when I want some lumpy of discretionary spend after the market has gone up, I already believe I have more than enough.

AlohaJoe
Posts: 4846
Joined: Mon Nov 26, 2007 2:00 pm
Location: Saigon, Vietnam

Re: Withdrawal Plan. McClung or Allocation Bands?

Post by AlohaJoe » Thu Oct 17, 2019 12:26 am

Uncorrelated wrote:
Wed Oct 16, 2019 2:52 am
It has been mathematically proven that if you cannot predict the market, then a constant allocation that is frequent rebalanced is the best possible asset allocation.
I'm not even a fan of glidepaths but I feel that you are overstating the case against them. Can you provide a link to this proof? I find the claim rather hard to believe given that financial economists have no consensus on what a realistic utility function looks like, which would be a precondition for determining "best".
Uncorrelated wrote:
Wed Oct 16, 2019 1:10 pm
See this paper, theory 1 on page 119: An Introduction to Computational Finance Without Agonizing Pain
Ah, okay, so there's not actually a mathematical proof that frequent rebalancing is the best possible asset allocation. It says they have the same (i.e. constant is not better, it is simply not worse) expected value of terminal wealth given a simplified model with certain assumptions. More complicated models that account for stochastic mortality, declining human capital, hyperbolic risk aversion, time-varying consumption based on health status, spending shocks, bequests, and so on don't always find the same result.

Uncorrelated
Posts: 105
Joined: Sun Oct 13, 2019 3:16 pm

Re: Withdrawal Plan. McClung or Allocation Bands?

Post by Uncorrelated » Thu Oct 17, 2019 3:03 am

2pedals wrote:
Wed Oct 16, 2019 11:41 pm
Whether a stratege is "active" of "passive" is unrelated to market timing.

If you have an asset allocation of 50/50, then you are taking a certain amount of risk for a certain amount of reward. If market timing is impossible, then the risk and reward of this particular allocation is constant over time.

Suppose the next day, stocks go up and bonds go down. Your net worth doesn't change, but your 50/50 asset allocation has changed to a 60/40 asset. You are now taking a different risk with different rewards than before. If you accept that, then you are irrational.
It is just a rule that says to sell some stocks if they go up.
That is just a roundabout way of predicting the market. i.e. market timing.
A constant allocation strategy is based an implicit bet against diverging and poor stock asset performance for the duration of your life or assets. If while in the duration of your retirement stocks under perform relative to bonds you will be buying stocks while at the same withdrawing money for retirement. A constant allocation strategy if done during a diverging duration will perform worse than a buy and hold strategy.
A buy & hold strategy with initial asset allocation of 50/50 outperforms a constant allocation with the same asset allocation (higher expected value and more risk). Of course it does, it takes on more risk. But there exists a constant allocation (say, 70/30) that offers the same expected value and lower risk than a buy & hold strategy.
My initial asset allocation was set based on how much assets I feel comfortable placing in the risk bucket for the duration of my retirement.
Excellent observation. Suppose that you retire today with two buckets: stocks and bonds. 50% in each.

Then the next day, stock go up and bonds go down. You end up with 60% in your stocks bucket and 40% in your bonds bucket, but the same net worth.

Why are you suddenly comfortable taking more risk than yesterday?

Buckets are mental gymnastics. Your risk is determined by your total asset allocation, not by your buckets.

AlohaJoe wrote:
Thu Oct 17, 2019 12:26 am
Ah, okay, so there's not actually a mathematical proof that frequent rebalancing is the best possible asset allocation. It says they have the same (i.e. constant is not better, it is simply not worse) expected value of terminal wealth given a simplified model with certain assumptions. More complicated models that account for stochastic mortality, declining human capital, hyperbolic risk aversion, time-varying consumption based on health status, spending shocks, bequests, and so on don't always find the same result.
The same expected value and lower risk, or higher expected value with the same risk.

if your risk tolerance is time-varying or your utility function is not isoelastic, then a constant allocation is not optimal. Good catch, I should have mentioned that. Maybe prime harvesting works because it approximates the optimal solution of a more complicated model. But there are just too many things wrong with it to take seriously.

Uncorrelated
Posts: 105
Joined: Sun Oct 13, 2019 3:16 pm

Re: Withdrawal Plan. McClung or Allocation Bands?

Post by Uncorrelated » Mon Oct 28, 2019 3:40 pm

As some of you have already seen, I published an analysis of optimal withdrawal strategies here:

viewtopic.php?f=10&t=293469

This may be of interest to some of the readers here.

Post Reply