Guyton-Klinger Portfolio Management Rule Confusion

Non-investing personal finance issues including insurance, credit, real estate, taxes, employment and legal issues such as trusts and wills
Post Reply
gasserma
Posts: 3
Joined: Sun Jun 11, 2017 5:12 pm

Guyton-Klinger Portfolio Management Rule Confusion

Post by gasserma » Fri Jun 16, 2017 6:31 pm

Hi Folks,

I have a question about an interpretation of the Guyton-Klinger portfolion management rule.

Here is the paper in question: http://cornerstonewealthadvisors.com/wp ... rticle.pdf
That pdf isn't ctrl+f'able so searching through it is hard. If anyone has a better link, that would be awesome.

The question I have is with regards to the definition of the portfolion management rule (PMR) that can be found at the bottom of the center column of page 2 of the pdf (which is labeled as page 51 of the excerpt).

To quote (and this isn't a copy pase given the nature of the pdf, so I may have made a typo copying it):

"The portfolio management rule (PMR) determines the source(s) of each year's withdrawal.
-Following years where an asset class has a positive return that produced a weighting exceeding its target allocation, the excess allocation is sold and the proceeds invested in cast to meet future withdrawal requirements.
-Portfolio withdrawals are funded in the following order: (1) overweighting in equity asset classes from the prior year-end, (2) overweighting in fixed income from the prior year-end, (3) cash, (4) withdrawals from remaining fixed income assets, (5) withdrawals from remaining equity assets in order of the prior year's performance."

Okay, now to my question/confusion...

This seems very brittle in its wording, particularly the first bullet. Imagine I have a hypothetical portfolio of two assets, each accounting for 50% allocation of my portfolio, which total $200 (so $100 of each of the following).
1. VSTAX
2. Poopcoin (a hypothetical tracker of an awful cryptocurrency)

On the first year of my retirement the total market is pretty benign, 1% growth. Poopcoin crashes though, -99%.

Now I have a portfolio that looks like:
1. VSTAX: $101
2. Poopcoin $1

And an allocation that looks like this:
1. VSTAX: 50%
2. Poopcoin: 50%

What do I do? As I read the the first bullet I need to sell $100 worth of vstax into cash.

That is obviously pretty dumb, and no real person would do that, but clarity here is important for programmatic back-testing. It would be bad if a bad asset could sabotage the whole strategy. There are other cases that fail too, like single asset class portfolios, etc.

I guess my questions then are, in order...
1. Did Guyton and Klinger address this ever and I missed it? I have heard that some of their future iterations simply omit the portfolio management rule, which is I guess might be the answer here, but that is a little unsatisfying. I really hope there is a definitive interpretation that isn't broken.
2. If not, how do you guys interpret this? It seems like a pretty broken piece of logic in the Guyton-Klinger strategy.

To give my own answers to 2:
I expect that the intent and a more workable version of this rule is that if your portfolio has not grown, you just rebalance. If your portfolio has grown, you take the minimum of [the net growth] and [the difference between the highest performing and lowest performing asset classes] and sell that to cash, rebalancing the rest.

WoodSpinner
Posts: 259
Joined: Mon Feb 27, 2017 1:15 pm

Re: Guyton-Klinger Portfolio Management Rule Confusion

Post by WoodSpinner » Fri Jun 16, 2017 11:53 pm

QQ...

Does the expense funding also include a subsequent rebalancing?

That is my approach if things get out of whack.

Your example is a pretty edge case, how does it work for your portfolio?

AlohaJoe
Posts: 2425
Joined: Mon Nov 26, 2007 2:00 pm
Location: Saigon, Vietnam

Re: Guyton-Klinger Portfolio Management Rule Confusion

Post by AlohaJoe » Sat Jun 17, 2017 7:55 am

gasserma wrote:On the first year of my retirement the total market is pretty benign, 1% growth. Poopcoin crashes though, -99%.

Now I have a portfolio that looks like:
1. VSTAX: $101
2. Poopcoin $1

And an allocation that looks like this:
1. VSTAX: 50%
2. Poopcoin: 50%

What do I do? As I read the the first bullet I need to sell $100 worth of vstax into cash.
I don't understand how you got to $100. Your portfolio (after the crash) is $102. VTSAX's target allocation is 50%. So the target allocation is $51. So the excess is $50, not $100.

So the decision rules don't work with ridiculous non-real world portfolios that an advisor like Guyton would never recommend to a client? Likely he would just shrug his shoulders and say his rules are intended for reasonable, well diversified portfolios held by real human beings. I guess there's some theoretical concern there but I'm doubtful about its real world impact. It wouldn't surprise me at all if Guyton's rules are not mathematically perfect.
That is obviously pretty dumb, and no real person would do that, but clarity here is important for programmatic back-testing. It would be bad if a bad asset could sabotage the whole strategy. There are other cases that fail too, like single asset class portfolios, etc.
Lots of people have done programmatic back-testing with Guyton-Klinger decision rules and never run into a non-theoretical problem with them when using real world data. But you can modify the rules any way you want; there's nothing saying they are written in stone.
I have heard that some of their future iterations simply omit the portfolio management rule, which is I guess might be the answer here, but that is a little unsatisfying.
Why is it unsatisfying? They did additional testing and found that it added no real world value. Getting rid of useless complications is always satisfying to me.
I expect that the intent and a more workable version of this rule is that if your portfolio has not grown, you just rebalance.
I don't think that's the intent. My recollection is that Guyton-Klinger never do "rebalancing" in the usual sense. You only indirectly rebalance by converting to cash. So it would be weird if they introduce rebalancing for this scenario.

livesoft
Posts: 56385
Joined: Thu Mar 01, 2007 8:00 pm

Re: Guyton-Klinger Portfolio Management Rule Confusion

Post by livesoft » Sat Jun 17, 2017 8:00 am

An important analysis of Guyton Klinger withdrawal rules is found in the EarlyRetirementNow series in parts 9 and 10:

https://earlyretirementnow.com/2017/02/ ... n-klinger/
This signature message sponsored by sscritic: Learn to fish.

User avatar
siamond
Posts: 3332
Joined: Mon May 28, 2012 5:50 am

Re: Guyton-Klinger Portfolio Management Rule Confusion

Post by siamond » Sat Jun 17, 2017 9:18 am

gasserma wrote:I have a question about an interpretation of the Guyton-Klinger portfolio management rule.

Here is the paper in question: http://cornerstonewealthadvisors.com/wp ... rticle.pdf
That pdf isn't ctrl+f'able so searching through it is hard. If anyone has a better link, that would be awesome.

The question I have is with regards to the definition of the portfolion management rule (PMR) that can be found at the bottom of the center column of page 2 of the pdf (which is labeled as page 51 of the excerpt).
The paper from Guyton-Klinger you quoted is quite old (2006). The authors were exploring ideas by then, some of them really good (e.g. guardrails and capital preservation / prosperity rules), others more dubious (e.g. PMR rule, inflation rule). I would suggest you focus on more recent writings of theirs, like this article from Jonathan Klinger in 2009.

As to the ERN discussion mentioned by livesoft in the previous post, personally, I found it quite misguided and missing the point. If you're curious, you can read some back & forth between myself and the author in the comments section.

We had a pretty lengthy exchange on Guyton-Klinger a little while ago on this forum. It got derailed a little bit, but you can probably find some useful posts in there: viewtopic.php?f=2&t=160073.

gasserma
Posts: 3
Joined: Sun Jun 11, 2017 5:12 pm

Re: Guyton-Klinger Portfolio Management Rule Confusion

Post by gasserma » Sat Jun 17, 2017 8:33 pm

Lots of people have done programmatic back-testing with Guyton-Klinger decision rules and never run into a non-theoretical problem with them when using real world data. But you can modify the rules any way you want; there's nothing saying they are written in stone.
The problem exists even without edge cases though. There are plenty of years where one asset class goes slightly up and one asset class goes slightly down. Its still unclear what they intended in that case, or rather, it is clear, but seems wrong.
They did additional testing and found that it added no real world value.
Where do they do this? That is essentially what I am looking for.
I don't understand how you got to $100. Your portfolio (after the crash) is $102. VTSAX's target allocation is 50%. So the target allocation is $51. So the excess is $50, not $100.
This may be what I am missing. Is that how to interpret the rules? I don't understand how your example goes though, at the end of it you have:
1. VSTAX: 50% allocation: $50.
2. Poopcoin: 50% allocation: $1.
3. Cash: $50.
That doesn't seem right to me.

gasserma
Posts: 3
Joined: Sun Jun 11, 2017 5:12 pm

Re: Guyton-Klinger Portfolio Management Rule Confusion

Post by gasserma » Sun Jun 18, 2017 11:49 am

The paper from Guyton-Klinger you quoted is quite old (2006). The authors were exploring ideas by then, some of them really good (e.g. guardrails and capital preservation / prosperity rules), others more dubious (e.g. PMR rule, inflation rule). I would suggest you focus on more recent writings of theirs, like this article from Jonathan Klinger in 2009.
Thanks, I think this ultimately is my answer...they just abandon the portfolio management rule. I still wish I could find some discussion on that. It seems pretty flimsy that they could just vanish it into thin air without a trace after publishing on it.

User avatar
One Ping
Posts: 256
Joined: Thu Sep 24, 2015 4:53 pm

Re: Guyton-Klinger Portfolio Management Rule Confusion

Post by One Ping » Sun Jun 18, 2017 12:36 pm

siamond wrote:I would suggest you focus on more recent writings of theirs, like this article from Jonathan Klinger in 2009.
William J. Klinger in 2007.
gasserma wrote:It seems pretty flimsy that they could just vanish it into thin air without a trace after publishing on it.
Why is that? That's what happens sometimes when further research is done ... like the 2007 Klinger paper. And it didn't vanish into thin air, he mentions it in his 2007 paper ...
Klinger, 2007 wrote:Earlier work by Guyton and Klinger (2006) used the Portfolio Management Decision Rule and did not rebalance assets each year. That rule does not materially affect the shape of a retirement profile and simple rebalancing was used for this research.
If after Bengen's early 90's seminal work on Sustainable Withdrawal Rates (SWR) no further research had been done we never would have had the Trinity Study, VPW, Prime Harvesting, or any number of other withdrawal strategies developed. Time marches on and knowledge continually evolves.
"Re-verify our range to target ... one ping only."

User avatar
siamond
Posts: 3332
Joined: Mon May 28, 2012 5:50 am

Re: Guyton-Klinger Portfolio Management Rule Confusion

Post by siamond » Sun Jun 18, 2017 12:50 pm

One Ping wrote:
siamond wrote:I would suggest you focus on more recent writings of theirs, like this article from Jonathan Klinger in 2009.
William J. Klinger in 2007.
LOL, you're perfectly right, thank you! I actually had in mind to quote another Klinger article from 2009, but couldn't find it anymore. Oh well. I do wish the authors had continued to publish a bit more on the topic, and notably made more of an effort to communicate in layman's terms. The guardrail mechanics are really very simple in truth, while those fairly academic articles somehow convey a sense of complexity that is really NOT at the core of the concept. I've looked at many withdrawal methods, and although not perfect, the Guyton-Klinger guardrail methodology is one of the few good ones imho...

User avatar
One Ping
Posts: 256
Joined: Thu Sep 24, 2015 4:53 pm

Re: Guyton-Klinger Portfolio Management Rule Confusion

Post by One Ping » Sun Jun 18, 2017 1:13 pm

siamond wrote:
One Ping wrote:
siamond wrote:I would suggest you focus on more recent writings of theirs, like this article from Jonathan Klinger in 2009.
William J. Klinger in 2007.
LOL, you're perfectly right, thank you! I actually had in mind to quote another Klinger article from 2009, but couldn't find it anymore. Oh well. I do wish the authors had continued to publish a bit more on the topic, and notably made more of an effort to communicate in layman's terms. The guardrail mechanics are really very simple in truth, while those fairly academic articles somehow convey a sense of complexity that is really NOT at the core of the concept. I've looked at many withdrawal methods, and although not perfect, the Guyton-Klinger guardrail methodology is one of the few good ones imho...
siamond, wasn't this one was it?

"Blanchett, David M., and Larry R. Frank Sr. 2009. “A Dynamic and Adaptive Approach to Distribution Planning and Monitoring.” Journal of Financial Planning 22 (4): 52–66."

Not Klinger, but on a similar subject and in 2009. I've generally always liked the way Blanchett approaches things, although sometimes not as rigorous as I would like.
"Re-verify our range to target ... one ping only."

User avatar
siamond
Posts: 3332
Joined: Mon May 28, 2012 5:50 am

Re: Guyton-Klinger Portfolio Management Rule Confusion

Post by siamond » Sun Jun 18, 2017 1:21 pm

One Ping wrote:siamond, wasn't this one was it?

"Blanchett, David M., and Larry R. Frank Sr. 2009. “A Dynamic and Adaptive Approach to Distribution Planning and Monitoring.” Journal of Financial Planning 22 (4): 52–66."

Not Klinger, but on a similar subject and in 2009. I've generally always liked the way Blanchett approaches things, although sometimes not as rigorous as I would like.
No, it was something else. Oh well, I shouldn't speak out of fuzzy memory! Those retired folks and their 'senior' moments, you know... You gave me a good laugh when catching not one, but TWO mistakes of mine in one single sentence! :wink:

Yes, Blanchett is a very interesting fellow, and I've read a lot of his literature, often not quite agreeing with his conclusions, but he's prolific and thought-provoking for sure.

User avatar
One Ping
Posts: 256
Joined: Thu Sep 24, 2015 4:53 pm

Re: Guyton-Klinger Portfolio Management Rule Confusion

Post by One Ping » Sun Jun 18, 2017 1:41 pm

siamond wrote:
One Ping wrote:siamond, wasn't this one was it?

"Blanchett, David M., and Larry R. Frank Sr. 2009. “A Dynamic and Adaptive Approach to Distribution Planning and Monitoring.” Journal of Financial Planning 22 (4): 52–66."

Not Klinger, but on a similar subject and in 2009. I've generally always liked the way Blanchett approaches things, although sometimes not as rigorous as I would like.
No, it was something else. Oh well, I shouldn't speak out of fuzzy memory! Those retired folks and their 'senior' moments, you know... You gave me a good laugh when catching not one, but TWO mistakes of mine in one single sentence! :wink:

Yes, Blanchett is a very interesting fellow, and I've read a lot of his literature, often not quite agreeing with his conclusions, but he's prolific and thought-provoking for sure.
Agree, on all counts.
"Re-verify our range to target ... one ping only."

Post Reply