[Book: Living off Your Money, by M. McClung (Prime Harvesting)]

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
Post Reply
User avatar
Leif
Posts: 2471
Joined: Wed Sep 19, 2007 4:15 pm

Re: [Book: Living off Your Money, by M. McClung (Prime Harvesting)]

Post by Leif » Sun Jul 17, 2016 9:52 am

Raybo wrote: You are "a big fan" of maintaining a set AA in retirement. I used to be, too. But, McClung presents data showing it isn't the best method for doing so. Unlike Kevin, I'm not sure my pile will last, so I want to make the best decision regarding harvesting my money. While I am open to other ideas, I find McClurg's work to be a strong argument for changing my approach.
+1. I'm at the same place.

AlohaJoe
Posts: 3684
Joined: Mon Nov 26, 2007 2:00 pm
Location: Saigon, Vietnam

Re: [Book: Living off Your Money, by M. McClung (Prime Harvesting)]

Post by AlohaJoe » Sun Jul 17, 2016 10:41 am

soboggled wrote:If you implement these strategies and have a spouse who isn't up on all this, better assign a computer as your trustee if you suffer any loss of mental capacity or predecease.
Isn't this the case for most strategies? You can always select a simpler, less effective, strategy by choosing to work for a few extra years.

soboggled
Posts: 901
Joined: Mon Jun 27, 2016 10:26 am

Re: [Book: Living off Your Money, by M. McClung (Prime Harvesting)]

Post by soboggled » Sun Jul 17, 2016 11:54 am

AlohaJoe wrote:
soboggled wrote:If you implement these strategies and have a spouse who isn't up on all this, better assign a computer as your trustee if you suffer any loss of mental capacity or predecease.
Isn't this the case for most strategies? You can always select a simpler, less effective, strategy by choosing to work for a few extra years.
Selecting a simpler one and working longer may be a good idea in any case, since the scientific approach seems to require Prussian-like precision and a lot of faith in back-tested solutions. Since I am not convinced the solutions are very robust, they may be a starting point but I would build in a LOT of margin for error.

heyyou
Posts: 3095
Joined: Tue Feb 20, 2007 4:58 pm

Re: [Book: Living off Your Money, by M. McClung (Prime Harvesting)]

Post by heyyou » Sun Jul 17, 2016 12:40 pm

McClung has chosen the best practices of many different methods for each to slightly improve retirement portfolio spending. He steadily compares each step of the improvements to Bengen's SWR to show the differences. His Chapter 3 WD comparisons are the sources of the small improvements that he does choose. He constantly compares the WD rate method to the optimal historical rate that is only known in retrospect.

McClung uses data to support his multitude of small improvements
(1) variable WDs based on both the remaining portfolio value
(2) and the recent inflation rate
(3) starting with an initial WD rate that is tuned/tilted by an average of the known market valuation measures, and the intended retirement length
(4) WDing from the initial 45% of bonds first, until they are exhausted or replenished
(5) replenishing the bonds by selling off the 20% stock gain when the remaining 55% of equities has grown by 120% of their inflation adjusted retirement day value

He acknowledges retiree's fears by suggesting that the retiree have some, to all, necessary spending supplied by guaranteed income (GI) depending on what suits the retiree. His WD floor and ceiling are far below, and far above, the calculated WD amounts so those limits would only kick in during extreme markets but they do smooth the variable WDs at those times.

To stretch the 1926-2010 data, he wraps it forward, starting over at 1926. The 30 and 40 year retirements of the year 2000 retiree start with the 2000 crash and recovery, then 2008 crash with recovery to 2010, then to 1926 before the 1929 crash occurs to the same retiree halfway through retirement. He also runs simulations with long term returns of specific percentages less than their historical averages. He trusts his results since they are tested to worse than historical sequences, but are still using relative returns with the momentum and correlations that would be lost with random Monte Carlo simulations.

(3) For the retirement day stock market valuation level, he takes the three known measurements (Shiller's and two others), divides each into segments of 1 through 4, then averages the current segment values to get a single number to describe the relative, current market valuation. The valuation chart is not precise, it is relative, as in, from 2 to 3 do this, from 3 to 3.5 take out this fraction of a percentage less, at 4 take out even less.

That is how he resolves the problem of the 1999 and 2000 twin retirees having such different portfolio values on retirement day. High stock valuations tilt the initial WD % lower, and lower valuations tilt it higher, then both retirees continue variable WDing based on their current portfolio values. The method helps portfolio longevity. He points out that a WD of 3.6% from an overvalued portfolio is not dissimilar income to 4.8% initially from a recently crashed portfolio when you are taking variable WDs.

The most noticeable use of supposedly three digit accuracy was when a surviving portfolio had numerous 4.00% annual withdrawals in the first decade of a 30 or 40 year, worst case retirement that in retrospect only supported a maximum of 4.01% WD average over the entire period. McClung commented that the .01% not spent annually during the initial decade had compounded into enough to help the portfolio during its final few years.

jjface
Posts: 2549
Joined: Thu Mar 19, 2015 6:18 pm

Re: [Book: Living off Your Money, by M. McClung (Prime Harvesting)]

Post by jjface » Sun Jul 17, 2016 2:08 pm

Raybo wrote:
jjface wrote: They both had the same portfolio value,mix, lifespan etc.
They aren't the same age and don't have the same lifespan. As long as we are creating straw men, assume we know that they will both die on the same day. Then, person 1 has had to fund 1 more year of retirement. Why should any method treat these as equivalent?
Well actually most harvesting methods ignore lifespan in their methodology. There is no planned end point.
In any case that one year of retirement extra is irrelevant as the portfolios are assumed equal after the first year of the earlier retiree to keep everything the same.
Raybo wrote:
jjface wrote: The method should at least take into account that person 2 started after equities crashed so the starting balance reference point is distorted. Similarly for person 1.
Person 1's AA has gone from roughly 65/35 to 55/45, while person 2 starts with 55/45. Why should two people with different starting AAs, though the same starting totals, invest the same?
Precisely because they have the exact same portfolio and allocation at any given time point before year 2. When person 1 retires with 65/35 person 2 has 65/35 (and the same portfolio balance). When person 2 retires with 55/45 then person 1 also has 55/45 (and the same balance). Yet they both have different references points as to what constitutes excess stock gains. That is surely illogical.

That is my problem with this method -there is an arbitrary lock in of the starting value.

The method does not take into account the cycle of the markets. For the one who retires near the bottom of a crash the method treats the initial few years of recovery as a large excessive stock gain when in fact it is just a recovery of prior losses. Similarly for someone who retires near the peak before a crash the method assumes that the excess gains are still part of the recovery and clings onto stocks longer (which is probably a good thing but is unintentional and adds extra risk). Surely any method like this one should look before the start point as well rather than just taking the start point as the reference point for the "normal level" of the stock market.
Raybo wrote: Under the constant AA in retirement scenario, person 1 would rebalance to 65/35 by selling bonds to both fund his/her yearly withdrawal and to move his/her stock amount up to $715k. Person 2 would simply sell stocks and bonds in the same 55/45 percentage to fund his/her withdrawal. On January 2, after all these changes have been made, persons 1 and 2 are still in completely different situations.

You appear to be suggesting that person 1 and 2 should do the same thing on January 1, which means you think the best move for person 1 is to change his/her AA as a result of what the market returns in the first year of person 1's retirement. This is the same tactical asset allocation that you say is a bad idea and should be avoided. You can't have it both ways
I was not comparing the method to another so the point is not related to what I was saying. I was trying to show how the one method under two virtually identical scenarios deviates in application when it shouldn't.
Raybo wrote: Your example doesn't show what you purport it to show.
Probably not clearly but you have misunderstood some of it. I hope the above makes more sense.
Last edited by jjface on Sun Jul 17, 2016 2:23 pm, edited 5 times in total.

jjface
Posts: 2549
Joined: Thu Mar 19, 2015 6:18 pm

Re: [Book: Living off Your Money, by M. McClung (Prime Harvesting)]

Post by jjface » Sun Jul 17, 2016 2:13 pm

I am not saying it is a bad method just that I think that those thinking about using this method should

a) be prepared to take the extra risk involved of varying the asset allocation and in particular the chances of periods of holding a high stock allocation in retirement. One may find they take on more risk than they can handle and not be able to sleep at night.

b) understand a spouse may not be able to follow it as easily in the event of their death. Changing the method part way though a couple's retirement may cause issues.

c) realize that their start value may not adequately reflect the normal level of the stock market and so the reference point for excess gains may not be appropriate in extreme cases.

User avatar
Raybo
Posts: 1659
Joined: Tue Feb 20, 2007 11:02 am
Location: San Francisco
Contact:

Re: [Book: Living off Your Money, by M. McClung (Prime Harvesting)]

Post by Raybo » Sun Jul 17, 2016 4:05 pm

jjface wrote: Well actually most harvesting methods ignore lifespan in their methodology. There is no planned end point.
Do you have any reference that supports this statement? Every thing I've read about retirement spending is based on a retirement of a certain length. Surely, the needs of a 40 year retirement would be different than one of 30 or 50 years. I'd suggest that the best plan for a retirement of 30 years would be different than one of 29 years.
you have misunderstood some of it.
So, let me try to understand the test case you are proposing.

There are two people #1 and #2 separated by one year of age. At time 0, both have the same AAs and account balances (and the exact same investments). #1 decides to retire at time 0 with a 65/35 AA and $1.5MM portfolio. #2, with the same portfolio, decides to work another year but doesn't add anything to his/her retirement investments so, after-taxes, #2 earns only what is spent in the year.

In that year, the market crashes and both people end the year with the same $1.1MM portfolio now allocated 55/45. #2 decides to retire at the start of year 2.

Is this the scenario you are putting forward? Are there more ifs, ands, or buts, you'd like to add?

Assuming I have this scenario correct, it makes sense to me that these two people would make different decisions about their retirement plans going forward.

You say it makes no sense to you that they would do so. It looks like we disagree. Oh well.

Further, you keep making the same point. I understand that you don't like retirement plans "with a memory." Continuing to assert that there is something wrong with such retirement plans without doing any research is your right. But, don't expect those of us who have been doing further research to necessarily agree with you.
No matter how long the hill, if you keep pedaling you'll eventually get up to the top.

User avatar
Raybo
Posts: 1659
Joined: Tue Feb 20, 2007 11:02 am
Location: San Francisco
Contact:

Re: [Book: Living off Your Money, by M. McClung (Prime Harvesting)]

Post by Raybo » Sun Jul 17, 2016 4:15 pm

jjface wrote:I am not saying it is a bad method just that I think that those thinking about using this method should

a) be prepared to take the extra risk involved of varying the asset allocation and in particular the chances of periods of holding a high stock allocation in retirement. One may find they take on more risk than they can handle and not be able to sleep at night.
How do you define "extra risk?" Keep in mind that the return sequence that results in high stock allocations are those where the stock market value has either fallen or not grown for longer lengths of time. This is precisely the time when the risk of holding more stocks should be less risky.
b) understand a spouse may not be able to follow it as easily in the event of their death. Changing the method part way though a couple's retirement may cause issues.
Is the issue the complexity of the withdrawal method or the fact that the non-financial spouse might have trouble following anything but the simplest retirement plan? Is annual rebalancing all that much simpler than always take from bonds and each year make a calculation to see if you should sell stocks and buy bonds?
c) realize that their start value may not adequately reflect the normal level of the stock market and so the reference point for excess gains may not be appropriate in extreme cases.
Huh?
No matter how long the hill, if you keep pedaling you'll eventually get up to the top.

jjface
Posts: 2549
Joined: Thu Mar 19, 2015 6:18 pm

Re: [Book: Living off Your Money, by M. McClung (Prime Harvesting)]

Post by jjface » Sun Jul 17, 2016 6:02 pm

Forget the other example. Let me try to simplify.

$1m equities in year 1 with $660k bonds. crash happens so equities now worth $500k. You rebalance to maybe $720k equities 480k bonds. Guy is laid off. Starts prime harvesting. Starting equity balance is $720k. Market recovers to $920k. Prime harvesting says sell stocks. Does that make sense when a year before retirement he had $1m in equities? Seems to go against the never sell stocks low goal to me. Prime harvesting misinterpreted a recovery for excessive stock growth.

Maybe it doesn't make that much of a difference though - i don't know.
Last edited by jjface on Sun Jul 17, 2016 7:02 pm, edited 2 times in total.

drkathryn
Posts: 111
Joined: Fri Jul 01, 2016 2:34 pm

Re: [Book: Living off Your Money, by M. McClung (Prime Harvesting)]

Post by drkathryn » Sun Jul 17, 2016 6:21 pm

It might make sense if one need to liquidate for any sensible reason.

BTW: Be sure to include the H in his name ir you will find books like this one: The Thief Who Spat In Luck's Good Eye (The Amra Thetys Series) (Volume 2) :annoyed

jjface
Posts: 2549
Joined: Thu Mar 19, 2015 6:18 pm

Re: [Book: Living off Your Money, by M. McClung (Prime Harvesting)]

Post by jjface » Sun Jul 17, 2016 7:55 pm

Raybo wrote:
jjface wrote: Well actually most harvesting methods ignore lifespan in their methodology. There is no planned end point.
Do you have any reference that supports this statement? Every thing I've read about retirement spending is based on a retirement of a certain length. Surely, the needs of a 40 year retirement would be different than one of 30 or 50 years. I'd suggest that the best plan for a retirement of 30 years would be different than one of 29 years.
No method can accurately predict how long I will live. Most make assumptions to cover worse case scenarios based on past data and assume a certain length of retirement. No method will be as precise to differentiate to the level you suggest as quite simply no one knows if they will live 29/30 or how ever many years.
Raybo wrote: Further, you keep making the same point. I understand that you don't like retirement plans "with a memory." Continuing to assert that there is something wrong with such retirement plans without doing any research is your right. But, don't expect those of us who have been doing further research to necessarily agree with you.
I was trying to make myself clearer since you did not understand what I was trying to say. I am certainly willing to listen. Please do respond with constructive comments that help the other party in the discussion rather than belittling any comments they make if they are contrary to your own.
Raybo wrote:How do you define "extra risk?"
The usual related to holding more equities and especially 100%. Plus the emotional issues related to the volatility.
Raybo wrote:Is the issue the complexity of the withdrawal method or the fact that the non-financial spouse might have trouble following anything but the simplest retirement plan? Is annual rebalancing all that much simpler than always take from bonds and each year make a calculation to see if you should sell stocks and buy bonds?
One would need to read and understand the book and understand how to use the spreadsheet and why it is doing what it is doing. Whilst you and your spouse may be comfortable with that I am sure there are plenty who would not be. Other methods can be very simple eg rebalancing can be practiced through an all in one fund.

heyyou
Posts: 3095
Joined: Tue Feb 20, 2007 4:58 pm

Re: [Book: Living off Your Money, by M. McClung (Prime Harvesting)]

Post by heyyou » Sun Jul 17, 2016 8:38 pm

McClung speaks of big risks like lack of income in retirement, and the need to tolerate small risks. He would consider a variable stock allocation in retirement as a small risk worth taking for the benefit of letting the stocks grow longer.

Your equity allocation % would grow slowly as the bonds are depleted during poor stock markets and your WDs based on portfolio value would be smaller during that time.

In a rising equity market, equity gains beyond the original level, would be sold off during that period to replenish the bonds allocation, so the retiree would not be entering an over-priced stock crash with an exceptionally high stock allocation.

As mentioned above, he suggests that the retiree have necessary expenses mostly paid from guaranteed income sources, or by income from the low floor of the portfolio WDs. Thus the variable part of the income would be for discretionary spending.

For the twins who separately retired just before and just after a stock market crash, the WDs based on recent portfolio value would be similar after the first couple of years.

If someone prefers buckets, spending from bonds first, with rebalancing determined by stock gains, would fit that description.

The author has addressed most of the obvious questions somewhere through the text, but no one has had the book long enough to be able to quickly turn to his remarks on specific questions.

AlohaJoe
Posts: 3684
Joined: Mon Nov 26, 2007 2:00 pm
Location: Saigon, Vietnam

Re: [Book: Living off Your Money, by M. McClung (Prime Harvesting)]

Post by AlohaJoe » Sun Jul 17, 2016 8:57 pm

jjface wrote:Maybe it doesn't make that much of a difference though - i don't know.
When I did a quick calculation, a 2007 and a 2008 retiree both waited until 2013 to sell stocks. That could have been sensitive to exact allocations and dates, however, so it is hardly dispositive.

Kitces's answer for the "timing paradox" is that you need to take into account valuations. (IIRC the argument goes something like a 60/40 allocation at high valuations is riskier than a 60/40 allocation a low valuations, so you're not really holding the same risk-adjusted portfolio anyway.) Prime Harvesting sorta, kinda indirectly (but not really) takes valuations into account; but I agree there is room for improvement in this area, even if it just to keep everything about Prime Harvesting the same but show a stronger linkage around valuations. Obviously, if you don't believe in valuations then Kitces's theory won't hold much water.

I haven't heard of other suggestions to resolve the timing paradox; I have a feeling that it is something that bothers people in a theoretical mathematical sense but not much with real-world advisors, so it probably doesn't get a whole lot of attention.

User avatar
siamond
Posts: 4155
Joined: Mon May 28, 2012 5:50 am

Re: [Book: Living off Your Money, by M. McClung (Prime Harvesting)]

Post by siamond » Mon Jul 18, 2016 5:14 pm

jjface wrote:Forget the other example. Let me try to simplify.

$1m equities in year 1 with $660k bonds. crash happens so equities now worth $500k. You rebalance to maybe $720k equities 480k bonds. Guy is laid off. Starts prime harvesting. Starting equity balance is $720k. Market recovers to $920k. Prime harvesting says sell stocks. Does that make sense when a year before retirement he had $1m in equities? Seems to go against the never sell stocks low goal to me. Prime harvesting misinterpreted a recovery for excessive stock growth.

Maybe it doesn't make that much of a difference though - i don't know.
Yes, you are absolutely correct, starting from an initial value (stock assets at T0), and using such number to calibrate the Prime Harvesting algorithm for the entire retirement period doesn't seem quite right if valuations aren't taken in account. Somebody retiring at the top of the 2000 craze might have to wait quite a while before getting to the 120% threshold... Personally, if I were to use Prime Harvesting, I'd try to make the initial value somewhat valuation-neutral (e.g. divide by a valuation multiple derived from CAPE, Tobin's Q, or something like that). Sure, such multiplier is itself a fuzzy number, but I don't think we need a lot of precision, just a coarse adjustment to avoid side-effects of retiring at a peak or a through.

As a side note, the (valuation-neutral) initial value may have to be adjusted down the road. One may receive a lump sum (e.g. inheritance), or withdraw less than expected or even contribute to the portfolio (e.g. occasional income from a side gig), etc. This could make things more complex, notably for early retirees. Kind of akin to the challenges of tracking a cost basis.

User avatar
Raybo
Posts: 1659
Joined: Tue Feb 20, 2007 11:02 am
Location: San Francisco
Contact:

Re: [Book: Living off Your Money, by M. McClung (Prime Harvesting)]

Post by Raybo » Mon Jul 18, 2016 8:07 pm

In Chapter 9, McClung attempts to factor in market valuation not by changing the initial stock value but by adjusting the initial withdrawal percentage.
No matter how long the hill, if you keep pedaling you'll eventually get up to the top.

User avatar
siamond
Posts: 4155
Joined: Mon May 28, 2012 5:50 am

Re: [Book: Living off Your Money, by M. McClung (Prime Harvesting)]

Post by siamond » Mon Jul 18, 2016 9:05 pm

Raybo wrote:In Chapter 9, McClung attempts to factor in market valuation not by changing the initial stock value but by adjusting the initial withdrawal percentage.
Yes, I am aware, but this is orthogonal. McClung tries hard to separate the 'harvesting' technique from the withdrawal method. The latter tells you how much to withdraw, the former tells you how to withdraw (while possibly varying your AA). I am not quite bought in making such distinction, nor in giving up the idea of a fixed AA, but it is certainly interesting to think along those lines.

EDIT: actually, although Chapter 9 does focus on the withdrawal method (and initial rate), the author quickly comes back to the harvesting technique on p236, although his suggestion is to choose between Prime and Alternate-Prime based on initial valuations. Er... that's weird.

User avatar
siamond
Posts: 4155
Joined: Mon May 28, 2012 5:50 am

Re: [Book: Living off Your Money, by M. McClung (Prime Harvesting)]

Post by siamond » Wed Jul 20, 2016 5:37 pm

I finally took the time to analyze Prime Harvesting in more depth. I didn't rush to it as I had quite some doubts in my mind, this seemed like a mix of bucketing and TAA (Tactical Asset Allocation), and I have yet to see a bucket or a TAA scheme that makes much sense. Plus I agree with longinvest that a strong 'memory effect' is an undesirable property (besides short-term smoothing), so the entire algorithm being calibrated by the initial portfolio value doesn't seem quite right. Before I go further, let me reiterate that I really enjoyed the book, for its mix of creativity and thorough research. This doesn't mean I agree with all conclusions though! :wink:

Here is the basis of my quick backtesting study:
- 40 years of retirement
- starting portfolio and AA: $1M, invested at 60/40 (US Total-Market; US Total-Bonds)
- Simba historical numbers (1972+), mixed with SBBI historical numbers (pre-1972)
- no fixed income (SSA, Pension, SPIA, whatever), only portfolio withdrawals
- you can click on every image in this post and following, to see a bigger chart
- Prime Harvesting as defined by the author, 120% threshold to replenish bonds by 20% of stocks.
- Prime Harvesting: blue line; classic annual rebalancing (after withdrawal): orange line; all $$ numbers are inflation-adjusted.

First, I used the (very naive) constant withdrawal method, 4% of the initial portfolio, inflation-adjusted over the years. Certainly a bad withdrawal method, but this helps to focus on the 'harvesting' side of things to begin with. Here is the outcome for cycles starting in 1955 (kind of average cycle), 1965 (ouch! cycle), 1975 (very rosy cycle). I didn't show the spend trajectory, as this is $40K (inflation-adjusted) all along in all cases.

1955: Stocks AA varies between 40% and 80%, without any meaningful effect on the portfolio trajectory
1965: the AA jumps high fast, is somewhat reset by the oil crisis, then back to 100% until the whole thing dies (admittedly slightly later than classic rebalancing, although starting in 1966 would inverse the order). This scenario is VERY problematic if you ask me.
1975: as the market roars in the 80s/90s, the AA drops a lot. One could argue this is fine, and helps sleeping at night in 2000 and 2009, although I don't find the point terribly convincing.

Image

User avatar
siamond
Posts: 4155
Joined: Mon May 28, 2012 5:50 am

Re: [Book: Living off Your Money, by M. McClung (Prime Harvesting)]

Post by siamond » Wed Jul 20, 2016 5:57 pm

What happened in the 1965 cycle? Two things. First, valuations were high by then (CAPE at 23/24 - before the stock buybacks days), making the initial portfolio multiplied by 120% a really high ceiling to reach, so bonds quickly got depleted, and the stocks AA went really high... And then the oil crisis stuck, hitting the portfolio really hard. By then, the $40k withdrawals were too much, and the portfolio went in a downward spiral, during which the AA was stuck at 100% stocks because the initial portfolio value times 120% was just unattainable. Note that with the 1975 cycle, we had kind of the reverse effect, the market roars, and the algorithm keeps adding bonds to the AA, certainly too much for my taste.

So what if we follow the suggestion I made a few posts before, and adjust the initial portfolio by a multiplier derived from a reasonable valuation metric? Let's pick the Shiller CAPE divided by the average CAPE from 1881 to the current year (to avoid ex post reasonings) as such multiplier. This seems more reasonable to define an initial portfolio value to compare to. But most of you are starting to cringe, as valuation metrics are fuzzy and controversial. Still, let's give it a try.

Well, the 1965 cycle looks a bit better to begin with, keeping the AA in check. But the oil crisis is too severe, and the whole thing goes to hell in the same way as before. The 1975 cycle looks better to me, although I still shake my head at going down to 20% stocks, plus the whole trajectory no longer makes a difference with classic rebalancing. Oh, and the mild 1955 cycle doesn't look too good either. So much for my attempt to improve the algorithm... :shock:

Later tonight, I'll discuss more realistic scenarios, replacing the constant withdrawal method by a VPW derivative, while using Prime Harvesting (or not).

Image

User avatar
siamond
Posts: 4155
Joined: Mon May 28, 2012 5:50 am

Re: [Book: Living off Your Money, by M. McClung (Prime Harvesting)]

Post by siamond » Wed Jul 20, 2016 9:21 pm

EDIT: AlohaJoe made me realize in this post that the way I combined Prime Harvesting and VPW in my simulation below could be viewed as debatable, as I kept the fixed VPW rate computed from the initial AA all along (I didn't recompute it every year based on the changes of AA). Well, my model below is one possible choice, and not necessarily a bad one, so I'll leave this post as is, while pondering other possibilities.
-------


Let's give it a try with VPW now. I used the regular VPW with one small twist, I added a 10 years 'grace period' to the duration of the retirement (retirement scenarios span 40 years, but I used 50 years for the depletion years). This avoids strange side-effects at the end of the period, and makes the scenario more realistic for various reasons (bequest, LTC, etc). Then I did the very exact same type of comparison as with the constant-withdrawal method. I did NOT use a valuation multiplier here since it wasn't convincing enough in previous tests (well, I did try, but this didn't really change key conclusions). Here are the 1955/65/75 cycles.

For the 1955 cycle, Prime Harvesting doesn't add anything (except variations in AA - which, by the way, would be tough to handle with a sizable taxable account). For the 1965 cycle, Prime Harvesting appears better, but this is essentially a side-effect of the AA becoming very (overly?) aggressive. And the 1975 cycle, the reverse applies, although one might argue that the Prime Harvesting portfolio trajectory is slightly smoother. This isn't quite conclusive, to be fair.

Image

Now let's include the variable spending in the charts. The spending budget trajectory largely mirrors the portfolio trajectory (as expected with VPW). There is no meaningful smoothing due to the AA adjustments of the harvesting technique. This tells me that the Prime Harvesting properties which appeal to intuition (spend from bonds, don't buy more stocks, sell stocks when 'up') are probably mostly illusory, a behavioral clutch without much ground in reality. See the stats in the next post to reinforce that point.

Image
Last edited by siamond on Thu Jul 21, 2016 11:19 am, edited 2 times in total.

User avatar
LadyGeek
Site Admin
Posts: 47438
Joined: Sat Dec 20, 2008 5:34 pm
Location: Philadelphia
Contact:

Re: [Book: Living off Your Money, by M. McClung (Prime Harvesting)]

Post by LadyGeek » Wed Jul 20, 2016 9:34 pm

An introduction to your charts would be helpful. I believe you are providing three scenarios represented by 3 rows of charts.

- First row starts at year 1955
- Second row starts at 1965
- Third row starts at 1975
- "AA" is Asset Allocation.

What do you mean by "trajectory"? Is it the slope of the line (direction)?

What are the designators "/1p", "/1c", "/2p", "/2c" in your chart series titles? I don't see a "/3p", "/3c" for the third presentation.
Wiki To some, the glass is half full. To others, the glass is half empty. To an engineer, it's twice the size it needs to be.

User avatar
siamond
Posts: 4155
Joined: Mon May 28, 2012 5:50 am

Re: [Book: Living off Your Money, by M. McClung (Prime Harvesting)]

Post by siamond » Wed Jul 20, 2016 9:41 pm

Sure enough, three specific cycles (1955, 1965, 1975), as representative as they are, do not tell the entire story. For those of you looking for more synthetic metrics across a whole set of cycles, I ran cycles starting from 1926 to 1990 (reducing the duration of the cycle when needs be). The first table below defines various scenarios, the second table provides various synthetic metrics. Click to visualize a larger image. For those of you eager to dive at this level, this should be fairly self-explanatory (send me a PM if not). I also included the Alternate Prime Harvesting method while I was at it (P_PH_EqBd parameter set to zero).

The bottomline is that I do not see in those metrics any meaningful advantage of using Prime Harvesting, whatever one's interpretation of risk (volatility, loss of income, loss of value, etc) might be. Now the author of the book (Mr McClung) did see a clear advantage, but his way of assessing the quality of the method is centered on computing the Minimum Safe Withdrawal Rate (MSWR) while using a constant-dollar withdrawal method. This seems to me to be the primary weakness of the book, actually. The MSWR math seems pretty much meaningless for people using more sensible (variable) withdrawal methods, as the difference between the CDW and VPW charts in the past 2 posts should demonstrate. In addition, the MSWR math solely focuses on one single data point in history (the worst cycle), which adds to the lack of significance. When looking at actual cycles, and more meaningful metrics, I just don't see the point (while I would not want to wildly vary my AA in the way the harvesting method suggests).

Image

Image

User avatar
siamond
Posts: 4155
Joined: Mon May 28, 2012 5:50 am

Re: [Book: Living off Your Money, by M. McClung (Prime Harvesting)]

Post by siamond » Wed Jul 20, 2016 9:50 pm

LadyGeek wrote:An introduction to your charts would be helpful. I believe you are providing three scenarios represented by 3 rows of charts.

- First row starts at year 1955
- Second row starts at 1965
- Third row starts at 1975
- "AA" is Asset Allocation.
I was trying to not be too verbose... :wink: Yes, you got it right.
LadyGeek wrote:What do you mean by "trajectory"? Is it the slope of the line (direction)?
Let me take an example. The first row describes a retirement cycle, with a retiree gaining his freedom in 1955. The AA trajectory shows how the AA varies over the (40) years of happy retirement, by documenting the % of equities (stocks) in the AA for a given year (this stays constant on the orange line, cf. the classic rebalancing technique; while it varies quite a lot with Prime Harvesting). The portfolio (value) trajectory shows, in real dollars, how the portfolio value evolves, following the market vagaries, the changes of AA, and the withdrawals enabling the happy retiree to enjoy life. Clearer?
LadyGeek wrote:What are the designators "/1p", "/1c", "/2p", "/2c" in your chart series titles? I don't see a "/3p", "/3c" for the third presentation.
Those are references to the scenario names I used. Unless one wants to check the detailed stat tables in my last post, this isn't important to understand, but "1p" is constant-withdrawal with prime-harvesting, "1c" is constant-withdrawal with classic-rebalancing, "2p" is VPW with prime-harvesting, etc. There is no "3p" or "3c" because those are the same scenarios, I just showed the annual (variable) spend instead of the annual AA in one chart.

User avatar
Kevin M
Posts: 10002
Joined: Mon Jun 29, 2009 3:24 pm
Contact:

Re: [Book: Living off Your Money, by M. McClung (Prime Harvesting)]

Post by Kevin M » Wed Jul 20, 2016 10:20 pm

Very impressive analysis, siamond!

Kevin
Wiki ||.......|| Suggested format for Asking Portfolio Questions (edit original post)

User avatar
TimeRunner
Posts: 1372
Joined: Sat Dec 29, 2012 9:23 pm

Re: [Book: Living off Your Money, by M. McClung (Prime Harvesting)]

Post by TimeRunner » Wed Jul 20, 2016 10:24 pm

I think the 'method' could have been boiled down to a simple online website calculator, and could be incorporated as an alternative analysis in the VPW spreadsheet if there was sufficient interest. Thank you for saving most of us $60. "If it sounds too good to be true...." Thanks again. :beer
"What'd ya expect in an opera, a happy ending?" -Bugs Bunny ...:... One cannot enlighten the unconscious.

User avatar
Raybo
Posts: 1659
Joined: Tue Feb 20, 2007 11:02 am
Location: San Francisco
Contact:

Re: [Book: Living off Your Money, by M. McClung (Prime Harvesting)]

Post by Raybo » Wed Jul 20, 2016 10:54 pm

Siamond,

Thanks for the demonstration. It appears that your conclusion is that harvesting according to a pre-set AA is no worse than using Prime Harvesting and that you don't like the idea of letting your AA fluctuate, even up to 100%. As a result, you seem satisfied to reject McClung's work and stick with VPW and maintaining a constant AA through retirement.

Fair enough. But, keep in mind that maintaining a pre-set AA has its challenges, as well. In 2008-9, it meant selling bonds and buying stocks as they dove down. In that stretch, I sometimes hit rebalancing bands every week. This takes commitment and strong fortitude.

I was retired during 2008-9 and watched in mounting horror as my asset pile lost 1/3 of its total value as I kept rebalancing into the crash. In the end, I couldn't continue to do that and let my AA fall to 40/60 (stocks/bonds) where it sits today. I'm curious how Prime Harvesting would stack up against VPW and a constant AA for those who started retirement in 2000?
No matter how long the hill, if you keep pedaling you'll eventually get up to the top.

AlohaJoe
Posts: 3684
Joined: Mon Nov 26, 2007 2:00 pm
Location: Saigon, Vietnam

Re: [Book: Living off Your Money, by M. McClung (Prime Harvesting)]

Post by AlohaJoe » Wed Jul 20, 2016 10:59 pm

siamond wrote:1965: the AA jumps high fast, is somewhat reset by the oil crisis, then back to 100% until the whole thing dies (admittedly slightly later than classic rebalancing, although starting in 1966 would inverse the order). This scenario is VERY problematic if you ask me.
Could you elaborate a bit more on why you consider this scenario problematic? Prime Harvesting lasts 10% longer before portfolio exhaustion compared to Annual Rebalancing, which means it has a higher chance of providing funds. On the surface, that seems like a straightforward improvement.
1975: as the market roars in the 80s/90s, the AA drops a lot. One could argue this is fine, and helps sleeping at night in 2000 and 2009, although I don't find the point terribly convincing.
I see both of these as Swedroe's common line about "need to take risk" in action. In one case your need to take risk goes up, so you take more risk. In the other case, your need to take risk has gone down, so you take less.

(Disclaimer: I also have my reservations about Prime Harvesting; I'm trying to understand better what you find problematic to see if it is the same/different as my own reservations.)

jjface
Posts: 2549
Joined: Thu Mar 19, 2015 6:18 pm

Re: [Book: Living off Your Money, by M. McClung (Prime Harvesting)]

Post by jjface » Wed Jul 20, 2016 11:09 pm

Thanks for the analysis!

It is all very interesting and I can see some benefits from using the method but I just can't get used to the idea of varying my asset allocation so much either (especially when I would have spent 30-40 years not practicing tactical asset allocation through the accumulation phase). It would play with my nerves too much too. I imagine those who held 100% equities for a long period before retiring might feel more comfortable with the method and might find it more suited to their tastes. It could probably be improved with a min level of stocks as someone willing to go 100% equities is unlikely to want to drop so low as 20% equities unless they are very flexible.

Thanks for looking into the valuations too
Last edited by jjface on Thu Jul 21, 2016 1:10 am, edited 1 time in total.

AlohaJoe
Posts: 3684
Joined: Mon Nov 26, 2007 2:00 pm
Location: Saigon, Vietnam

Re: [Book: Living off Your Money, by M. McClung (Prime Harvesting)]

Post by AlohaJoe » Wed Jul 20, 2016 11:14 pm

siamond wrote:Let's give it a try with VPW now. I used the regular VPW with one small twist
Interesting, I've never tried to combine VPW with Prime Harvesting because of the complications. How did you solve them? VPW requires you to set your asset allocation (since it takes expected returns into account). And Prime Harvesting changes your asset allocation every year. So you have to recalculate the VPW table every single year.

The net result is what when you combine Prime Harvesting and VPW they ironically work against each other. Prime Harvesting increases your stock percentage during bad markets (since you are selling bonds and not replenishing them). And VPW sees your increasing stock percentage and tells you take an even bigger percent every year. For a retiree in the year 2000 using VPW with Annual Rebalancing will tell you to take 6.0% in the year 2010. But with Prime Harvesting it will tell you take to 6.4%.

I reckon that's almost completely the opposite of what most people would be doing, though.

AlohaJoe
Posts: 3684
Joined: Mon Nov 26, 2007 2:00 pm
Location: Saigon, Vietnam

Re: [Book: Living off Your Money, by M. McClung (Prime Harvesting)]

Post by AlohaJoe » Wed Jul 20, 2016 11:24 pm

Raybo wrote:I'm curious how Prime Harvesting would stack up against VPW and a constant AA for those who started retirement in 2000?
It depends on how you want to measure things. But just looking at portfolio values, Prime Harvesting did worse. That's because between 2003 and 2008 the markets didn't rise enough to trigger harvesting. Which means your stock percentage drifts up until it is around 80% right before the 2008 crash. And you'd be at 100% stocks today.

The net result is a bigger drop in 2008 and a bigger recovery afterwards. So, just looking at portfolio value at end of 2015, Prime Harvesting comes out ahead. But due to how we think about losses, I bet most people would focus on the bigger drop in 2008 and prefer the smoother ride to a smaller portfolio.

Image

User avatar
Raybo
Posts: 1659
Joined: Tue Feb 20, 2007 11:02 am
Location: San Francisco
Contact:

Re: [Book: Living off Your Money, by M. McClung (Prime Harvesting)]

Post by Raybo » Wed Jul 20, 2016 11:27 pm

I also am wondering what impact starting with 60% stocks has on your analysis? McClung suggests that 50/50 is the best AA to start with.

With $1MM portfolio, that means $500K in bonds, which at $50k/year would be 10 years, at least, before your AA got to 100% stocks. While there are starting years where this happened, there aren't any in the data where Prime Harvesting failed over 30 years.

Since you want to control stock percentage, I also wonder what impact setting the bond replenishment level to 100% of starting value instead of 120% might have on your analysis. McClung says that changing this stock selling threshold has only a small effect on the outcome.
No matter how long the hill, if you keep pedaling you'll eventually get up to the top.

User avatar
siamond
Posts: 4155
Joined: Mon May 28, 2012 5:50 am

Re: [Book: Living off Your Money, by M. McClung (Prime Harvesting)]

Post by siamond » Wed Jul 20, 2016 11:27 pm

TimeRunner wrote:I think the 'method' could have been boiled down to a simple online website calculator, and could be incorporated as an alternative analysis in the VPW spreadsheet if there was sufficient interest. Thank you for saving most of us $60. "If it sounds too good to be true...." Thanks again. :beer
Actually the author provides a simple-to-use spreadsheet with his book, with all primary recommendations implemented in there. And all I analyzed here is the Prime Harvesting method, there are many more topics discussed in the book. For me, this book is essentially great & innovative research, and one has to take it as such, with the caveats that it implies.

User avatar
siamond
Posts: 4155
Joined: Mon May 28, 2012 5:50 am

Re: [Book: Living off Your Money, by M. McClung (Prime Harvesting)]

Post by siamond » Wed Jul 20, 2016 11:44 pm

Raybo wrote:Siamond,

Thanks for the demonstration. It appears that your conclusion is that harvesting according to a pre-set AA is no worse than using Prime Harvesting and that you don't like the idea of letting your AA fluctuate, even up to 100%. As a result, you seem satisfied to reject McClung's work and stick with VPW and maintaining a constant AA through retirement.
Actually, I would not be philosophically opposed to let my AA fluctuate *to an extent* (I analyzed multiple flavors of Tactical Asset Allocation techniques in the past with an open mind, just didn't find anything worth it). But certainly not as wildly as Prime Harvesting does. But yes, you nicely summarized my thinking, I just don't see any clear value-added in this Prime Harvesting idea, while I do see significant issues:
- the AA can vary wildly
- it seems difficult (if not impossible) to implement with a large taxable account (can't change your AA on a dime without triggering capital gains)
- the calibration on the initial portfolio value is dubious at best, and displays way too much of a 'memory effect', with unfortunate side-effects
Raybo wrote:Fair enough. But, keep in mind that maintaining a pre-set AA has its challenges, as well. In 2008-9, it meant selling bonds and buying stocks as they dove down. In that stretch, I sometimes hit rebalancing bands every week. This takes commitment and strong fortitude.

I was retired during 2008-9 and watched in mounting horror as my asset pile lost 1/3 of its total value as I kept rebalancing into the crash. In the end, I couldn't continue to do that and let my AA fall to 40/60 (stocks/bonds) where it sits today. I'm curious how Prime Harvesting would stack up against VPW and a constant AA for those who started retirement in 2000?
Yes, I hear you about the psychological difficulties to manage a constant AA in times of crisis. I just don't see any factual value-added in trying to do any different though (what may make intuitive sense, e.g. not rebalance when things seem to be falling apart, just doesn't -or didn't- help). This being said, I suspect that NOT rebalancing for a year or two of deep crisis may not be overly consequential either.

I did ponder about the 2000 cycle: it doesn't make a lot of difference for the portfolio trajectory, while the AA quickly jumps to 100% stocks and stays stuck in there. Same problem as 1965 -well, worse!-, starting from a very overvalued portfolio makes the algorithm burn through bonds, then never replenish. Here my valuation trick would have been more helpful, bringing the algorithm to more reason, but then again, no value-added that I can see...

AlohaJoe
Posts: 3684
Joined: Mon Nov 26, 2007 2:00 pm
Location: Saigon, Vietnam

Re: [Book: Living off Your Money, by M. McClung (Prime Harvesting)]

Post by AlohaJoe » Wed Jul 20, 2016 11:50 pm

siamond wrote:When looking at actual cycles, and more meaningful metrics, I just don't see the point (while I would not want to wildly vary my AA in the way the harvesting method suggests).
I think you make a good point about "realistic withdrawal strategies" muting the results.

I had a hard time visualising the differences between the two from your table so I rewrote it as to help me understand better:

Image

Honestly, I was a bit surprised that the portfolio standard deviations were so close.

User avatar
siamond
Posts: 4155
Joined: Mon May 28, 2012 5:50 am

Re: [Book: Living off Your Money, by M. McClung (Prime Harvesting)]

Post by siamond » Thu Jul 21, 2016 12:12 am

AlohaJoe wrote:
Siamond wrote:1965: the AA jumps high fast, is somewhat reset by the oil crisis, then back to 100% until the whole thing dies (admittedly slightly later than classic rebalancing, although starting in 1966 would inverse the order). This scenario is VERY problematic if you ask me.
Could you elaborate a bit more on why you consider this scenario problematic? Prime Harvesting lasts 10% longer before portfolio exhaustion compared to Annual Rebalancing, which means it has a higher chance of providing funds. On the surface, that seems like a straightforward improvement.
First off, the 1966 cycle shows the very exact reverse behavior, Prime-Harvesting fails first. Well, if I got the math right, of course (my withdrawal-methods spreadsheet is getting complicated!). Fact is the algorithm is actually VERY sensitive to the initial-portfolio-value. Plus, as I tried to explain, this MSWR reasoning seems entirely irrelevant, nobody in their right mind will follow such constant-withdrawal technique down to the ground.

I think that 1965 clearly shows that the calibration on the initial portfolio-value (valuation-adjusted or not) is just not appropriate. Here we have a double-whammy, starting from a high valuation (hence a stock value threshold really high to reach), therefore you burn bonds and increase stocks, all that to fall on the knife of one of the worst crisis in history when you really wish you hadn't such a high % of stocks when the crisis strikes. And then you're stuck to 100% stocks forever (you'd probably stop taking $40k a year in real life, and somehow survive, but your portfolio is extremely unlikely to catch up with the 1965 heights). It might actually not be a bad decision to be at 100% stocks from a purely rational perspective, but that would be unbearable after living through such roller-coaster.
AlohaJoe wrote:
Siamond wrote:1975: as the market roars in the 80s/90s, the AA drops a lot. One could argue this is fine, and helps sleeping at night in 2000 and 2009, although I don't find the point terribly convincing.
I see both of these as Swedroe's common line about "need to take risk" in action. In one case your need to take risk goes up, so you take more risk. In the other case, your need to take risk has gone down, so you take less.
Yes, I can see such perspective which is exactly why I worded my sentence the way I did. Personally, this doesn't impress me because the 'orange' portfolio always stays above the 'blue' line, but that's just me, and I appreciate that other folks could see things differently here. This being said, in such situation, starting from a $1M portfolio and reaching heights like $3M, I'd guess that somebody following Larry's framework would change their (fixed) AA to more bonds, but based on milestones (e.g. $2M reached, $3M reached, or something like that), not based on a harvesting mechanism. I probably would.
AlohaJoe wrote:(Disclaimer: I also have my reservations about Prime Harvesting; I'm trying to understand better what you find problematic to see if it is the same/different as my own reservations.)
Thanks for being appreciative of my efforts and trying to understand my perspective. I tried to understand your posts before doing my own analysis, but the fact that you moved straight to the combination of Prime Harvesting and EM made it hard for me to unravel what was truly going on. In any case, by all means, please criticize my approach and/or conclusions as you see fit, this will make for a fruitful exchange. And maybe I did screw up the math somewhere...

User avatar
siamond
Posts: 4155
Joined: Mon May 28, 2012 5:50 am

Re: [Book: Living off Your Money, by M. McClung (Prime Harvesting)]

Post by siamond » Thu Jul 21, 2016 12:26 am

AlohaJoe wrote:
siamond wrote:Let's give it a try with VPW now. I used the regular VPW with one small twist
Interesting, I've never tried to combine VPW with Prime Harvesting because of the complications. How did you solve them? VPW requires you to set your asset allocation (since it takes expected returns into account). And Prime Harvesting changes your asset allocation every year. So you have to recalculate the VPW table every single year.
This is where having argued at length with longinvest about the pros and cons of VPW really helps! I know the inners of the beast! :D

VPW boils down to a very simple Excel formula: spend-budget = -PMT(expected-rate-of-returns, # of remaining depletion years, current-portfolio-value, 0, 1). And longinvest uses a very simple formula for the expected-rate, derived from the AA and the historical returns tracked by Credit Suisse. Hm, you actually made me realize I made a dubious choice though. I kept the expected-rate constant (based on the initial AA), I didn't recompute it every year based on the new AA. That's interesting... Ok, well, it's late now, but tomorrow, maybe I'll try to do it while taking in account the new AA every year.
AlohaJoe wrote:The net result is what when you combine Prime Harvesting and VPW they ironically work against each other. Prime Harvesting increases your stock percentage during bad markets (since you are selling bonds and not replenishing them). And VPW sees your increasing stock percentage and tells you take an even bigger percent every year. For a retiree in the year 2000 using VPW with Annual Rebalancing will tell you to take 6.0% in the year 2010. But with Prime Harvesting it will tell you take to 6.4%.

I reckon that's almost completely the opposite of what most people would be doing, though.
Yeah, well, this isn't what I modeled, as I just realized. But maybe I should have! :oops: I need to think a bit more about it though.

AlohaJoe
Posts: 3684
Joined: Mon Nov 26, 2007 2:00 pm
Location: Saigon, Vietnam

Re: [Book: Living off Your Money, by M. McClung (Prime Harvesting)]

Post by AlohaJoe » Thu Jul 21, 2016 1:20 am

siamond wrote:Yeah, well, this isn't what I modeled, as I just realized. But maybe I should have! :oops: I need to think a bit more about it though.
And to make things even worse, you need to use what someone in 1965 thought the expected returns were. Not what someone in 2016, with another 51 years of history, thinks the expected returns are.

The inclusion of expected returns is one reason why I don't like using VPW or Ken Steiner's approach in backtesting; it isn't impossible but it adds a layer of complexity and I'm lazy :)

AlohaJoe
Posts: 3684
Joined: Mon Nov 26, 2007 2:00 pm
Location: Saigon, Vietnam

Re: [Book: Living off Your Money, by M. McClung (Prime Harvesting)]

Post by AlohaJoe » Thu Jul 21, 2016 1:32 am

Raybo wrote:I also am wondering what impact starting with 60% stocks has on your analysis? McClung suggests that 50/50 is the best AA to start with.
This is a 2000-era retiree through the crash. The first chart is both start at 60/40. The second chart is Annual Rebalancing starts at 60/40 and Prime Harvesting starts at 50/50.

(Blue is annual rebalancing; Green is Prime)

Image

Image

It performs better. But we're also starting to drift further from an apples-to-apples comparison. Or are we? I don't actually know where to draw the line between fair comparisons and tilting the field toward one side or another. McClung clearly believes all of his pieces work well separately. But also presents it as a system where it all works together to make retirement more "efficient" for a retiree. So in some sense the fairest comparison might be to use everything -- starting stock percentages, portfolio construction, initial withdrawal rate tilt -- and compare that against the full package of status quo. Of course, then you lose sight of what is adding to outperformance (or underperformance) and what is a placebo.

longinvest
Posts: 2907
Joined: Sat Aug 11, 2012 8:44 am

Re: [Book: Living off Your Money, by M. McClung (Prime Harvesting)]

Post by longinvest » Thu Jul 21, 2016 7:55 am

I don't understand why most of the above backtests are done with aggressive portfolios (over 50% in stocks). Is a retiree's goal really to maximize withdrawals? Shouldn't it be, instead, to reduce anxiety with a relatively boring nominal withdrawal schedule and low nominal portfolio volatility, while generally keeping up with inflation?

There's a good reason the VPW backtesting spreadsheet displays both nominal and inflation-adjusted numbers. Even though CPI is a reliable long-term indicator of loss of purchase power, its year-on-year volatility is high, much higher than the impact of inflation on people who live in the real world and adjust their spending, in the short term, when prices go up. Example: Car prices shoot up; I'll repair the car instead of buying a new one. Furthermore, some retirees can be insulated from a major part of the CPI calculation; a retiree who owns his home free and clear could be pretty much insulated from an increase in rents.

It would be quite interesting to see the portfolio and withdrawal trajectory of 30/70, 40/60, and maybe 50/50 (stocks/bonds) constant AA portfolios vs Prime Harvesting in both nominal and inflation-adjusted term, and look at the result with the eye of a retiree who sees nominal numbers on his bank account and mutual fund statements.

The idea is that if prices go up, this year, and my nominal withdrawals (and portfolio balance) go up, but not as much as prices, I'll probably be happier than if prices didn't go up but my nominal withdrawals (and portfolio balance) went down. Yes, it is a behavioral bias, but it is real and affects many of us. As long as withdrawals are able to catch up with inflation soon enough, many of us are likely to feel good about a withdrawals increasing, even if slower than prices for a while.

For backtesting, year 1937 is a pretty challenging year for a 30/70 (stocks/bonds) portfolio. Here's a chart of nominal (blue) and inflation-adjusted (red) withdrawals (left) and portfolio balance (right) for a 35-year retirement (at age 65) using a portfolio of 30% US Stocks and 70% US Bonds*.

In inflation-adjusted term, the lowest withdrawal was $27,483 (the median, over 35 years, was $35,455). Notice how the nominal (blue) withdrawals were smoothly increasing in most years, and negative moves were very small, even though this is a plain unmodified VPW simulation (no smoothing whatsoever) with boring yearly rebalancing at the same time as withdrawals:

Image

For consistency with previous charts, on this thread, I did not include any amount of base income (Social Security, pension, or SPIA). This is unrealistic. In real life, there would be base income, and this would reduce the relative year-on-year variations even more.

Just a hunch: I suspect that Prime Harvesting might yield a higher average (and median?) withdrawal, but worse nominal portfolio single-year drops. The question is: Is the retiree appropriately "compensated" for this higher anxiety? Actually, is there any good reason for a retiree to assume a higher anxiety and lose sleep, just to get a few additional dollars that he might not even enjoy, as stress is a know factor for reducing life expectancy? :wink:

* US stocks returns are based on Shiller's data, US Bonds returns are based on a self-correcting simulation of a ladder-like fund which buys 10-years bonds and sells them 1 year prior to maturity, based on Shiller and FRED data (See: viewtopic.php?f=10&t=179425). The chart can be reproduced using the VPW backtesting spreadsheet.
Bogleheads investment philosophy | Lifelong Portfolio: 25% each of (domestic/international)stocks/(nominal/inflation-indexed)bonds | VCN/VXC/VLB/ZRR

User avatar
siamond
Posts: 4155
Joined: Mon May 28, 2012 5:50 am

Re: [Book: Living off Your Money, by M. McClung (Prime Harvesting)]

Post by siamond » Thu Jul 21, 2016 10:29 am

Raybo wrote:I also am wondering what impact starting with 60% stocks has on your analysis? McClung suggests that 50/50 is the best AA to start with.

With $1MM portfolio, that means $500K in bonds, which at $50k/year would be 10 years, at least, before your AA got to 100% stocks. While there are starting years where this happened, there aren't any in the data where Prime Harvesting failed over 30 years.
longinvest wrote:I don't understand why most of the above backtests are done with aggressive portfolios (over 50% in stocks). Is a retiree's goal really to maximize withdrawals? Shouldn't it be, instead, to reduce anxiety with a relatively boring nominal withdrawal schedule and low nominal portfolio volatility, while generally keeping up with inflation?
I ran my test with a 60/40 AA, because this is how McClung started his own analysis when comparing harvesting methods (see the book, p40). I used a simpler (US-only) AA than he did, because this seemed representative enough (truth is my spreadsheet is just not designed to let asset sub-classes vary in allocation!).

I was also very comfortable doing that because this is the historic recommendation of luminaries like Peter Bernstein, and a very reasonable choice imho. McClung rightfully makes the point multiple times that we shouldn't fret too much about high stock percentage in retirement (e.g. p72), and this has been my overarching conclusion when running numerous backtests on other topics. Focusing on 'loss of income' (over the entire retirement period) has to be much more important than focusing on '(temporary) loss of (portfolio) value', and this implies to keep an engine of growth in one's portfolio for long retirement periods (until one is very wealthy, and then harvesting/withdrawal methods become mostly irrelevant). I understand that other people might see otherwise, but well, that is probably a discussion which would side-track this thread, so maybe we should leave it as that.

Now would Prime Harvesting work better with 50/50 as a starting point? Well, a solid Harvesting method should display good behavior, irrespective of the starting AA, but still I gave it a quick test with constant-withdrawal, and yes, this makes the 1965 scenario a bit more reasonable (still goes to the ground, but at least the first part is more reasonable). The overall stats for the 1926-1990 set of cycles are also somewhat improved - actually, it's funny, as such AA fails more often than 60-40 (hint, hint), PH mitigates a bit better those really bad scenarios. Maybe this is the point I missed in my previous test, if you get an SWR closer to possible total-failure scenarios, PH starts behaving in a more satisfactory manner (heck, it was calibrated on MSWR, so this makes sense). But then how to do choose such SWR? And what if you're wrong?

Again, using a constant-withdrawal method and discussing MSWR at length is pretty pointless imho. Better see if the use of Prime Harvesting can improve the spend/portfolio trajectory when using a solid variable withdrawal method, VPW, Guyton-Klinger, McClung's EM or else.

PS. and I am not going to enter in a discussion about inflation representations and inflation management again... Longinvest and myself argued about this many times, we just don't convince each other at all on this matter, so better not derail this thread.

User avatar
TimeRunner
Posts: 1372
Joined: Sat Dec 29, 2012 9:23 pm

Re: [Book: Living off Your Money, by M. McClung (Prime Harvesting)]

Post by TimeRunner » Thu Jul 21, 2016 10:37 am

siamond wrote:
TimeRunner wrote:I think the 'method' could have been boiled down to a simple online website calculator, and could be incorporated as an alternative analysis in the VPW spreadsheet if there was sufficient interest. Thank you for saving most of us $60. "If it sounds too good to be true...." Thanks again. :beer
Actually the author provides a simple-to-use spreadsheet with his book, with all primary recommendations implemented in there. And all I analyzed here is the Prime Harvesting method, there are many more topics discussed in the book. For me, this book is essentially great & innovative research, and one has to take it as such, with the caveats that it implies.
I had looked at the spreadsheet, but didn't feel confident about the method, and I'm not quisitive enough to explore the book. I'll keep an eye on the thread from the grandstands and will probably stay off the field. Enjoy the journey. :beer
"What'd ya expect in an opera, a happy ending?" -Bugs Bunny ...:... One cannot enlighten the unconscious.

User avatar
siamond
Posts: 4155
Joined: Mon May 28, 2012 5:50 am

Re: [Book: Living off Your Money, by M. McClung (Prime Harvesting)]

Post by siamond » Thu Jul 21, 2016 12:06 pm

longinvest wrote:Just a hunch: I suspect that Prime Harvesting might yield a higher average (and median?) withdrawal, but worse nominal portfolio single-year drops. The question is: Is the retiree appropriately "compensated" for this higher anxiety? Actually, is there any good reason for a retiree to assume a higher anxiety and lose sleep, just to get a few additional dollars that he might not even enjoy, as stress is a know factor for reducing life expectancy? :wink:
With the way I combined VPW (keeping a fixed rate) and Prime Harvesting, the average withdrawal across all cycles I tested is indeed higher (hardly surprising as the average AA over all cycles is a tad more aggressive than the initial 60/40 AA). BUT. The average withdrawal is *lower* for the really rosy cycles (fair enough), and also *lower* for the really bad cycles (now that is troublesome). The 1966 cycle is the worst, due to the AA increase before 1973 and then the oil shock. Can't comment about one-year drops, I don't have it in my spreadsheet.

Even if you and I do not have the same definition of risk and primary drivers of anxiety, I wouldn't disagree with your conclusion, the Prime Harvesting ride remains unnecessarily wild imho - while its aim was supposed to relieve some of the anxiety (cf. sell bonds, don't buy stocks, etc).

EDIT: improved the wording.
Last edited by siamond on Thu Jul 21, 2016 1:24 pm, edited 1 time in total.

longinvest
Posts: 2907
Joined: Sat Aug 11, 2012 8:44 am

Re: [Book: Living off Your Money, by M. McClung (Prime Harvesting)]

Post by longinvest » Thu Jul 21, 2016 12:53 pm

siamond wrote: With the way I combined VPW (keeping a fixed rate) and Prime Harvesting, the average withdrawal across all cycles I tested is indeed higher (hardly surprising as the average AA is a tad more aggressive than the initial 60/40 AA). BUT. The average withdrawal is *lower* for the really rosy years (fair enough), and also *lower* for the really bad years (now that is troublesome).
Yes, this is surprising. Yet, when I look at your AA trajectory charts, I understand how that comes to be. By making the AA bond-heavy in good years, Prime Harvesting lowers returns in good years, and by increasing the stock allocation is low return years, it increases the portfolio's vulnerability to stock market crashes.

There's something difficult to beat about a good old fixed AA, adjusted to the investor's own volatility/fear-of-running-inflation/whatever tolerance.
siamond wrote: Even if you and I do not have the same definition of risk and primary drivers of anxiety, ...
I don't think that it is reasonable to expect everybody to share the same perspective on various risks. I think that we both agree that part of the risk assessment of a portfolio and withdrawal strategy is of a very personal nature. The investor's own circumstances, his beliefs, and his relative understanding of various investments will have a major impact on the choice of an AA and Investment Policy Statement (IPS, which covers contribution/withdrawal strategy, rebalancing method, and more).

But, once one chooses some specific definition or perspective for himself, it is important to stay coherent with it. That's why I expressed, earlier in this thread, that to me it would make no sense for an investor to maintain a fixed AA during accumulation, and then switch to some form of dynamic (or whatever you want to call Prime Harvesting's non-fixed) AA in retirement, without a change in beliefs.
Bogleheads investment philosophy | Lifelong Portfolio: 25% each of (domestic/international)stocks/(nominal/inflation-indexed)bonds | VCN/VXC/VLB/ZRR

User avatar
siamond
Posts: 4155
Joined: Mon May 28, 2012 5:50 am

Re: [Book: Living off Your Money, by M. McClung (Prime Harvesting)]

Post by siamond » Thu Jul 21, 2016 1:20 pm

longinvest wrote:But, once one chooses some specific definition or perspective for himself, it is important to stay coherent with it. That's why I expressed, earlier in this thread, that to me it would make no sense for an investor to maintain a fixed AA during accumulation, and then switch to some form of dynamic (or whatever you want to call Prime Harvesting's non-fixed) AA in retirement, without a change in beliefs.
One could argue that the dynamics in distribution/retirement are different enough from the dynamics in accumulation to justify *some* change in methodology, but I would tend to agree with you, here this is just way too much of a departure.

And I can't see how a derivative of Prime Harvesting could be used during accumulation (one needs to buy stocks from time to time during accumulation!).

User avatar
Kevin M
Posts: 10002
Joined: Mon Jun 29, 2009 3:24 pm
Contact:

Re: [Book: Living off Your Money, by M. McClung (Prime Harvesting)]

Post by Kevin M » Thu Jul 21, 2016 3:38 pm

siamond wrote: One could argue that the dynamics in distribution/retirement are different enough from the dynamics in accumulation to justify *some* change in methodology, but I would tend to agree with you, here this is just way too much of a departure.
One thing that clearly changes is that human capital decreases as you approach retirement, which tends to lower your ability to take risk. This could justify switching to more of a Liability Matching Portfolio / Risky Portfolio (LMP/RP) paradigm from the AA/rebalance paradigm that might be more appropriate during early-to-mid accumulation years.

The interesting thing is that any method that could put you 100% into stocks is about as far from an LMP/RP approach as you could get. I am much more persuaded by the cautious advice from Bill Bernstein, taking deep risk into account, than by a method based purely on historical results that probably don't provide enough data to have much statistical certainty in the conclusions.

Kevin
Wiki ||.......|| Suggested format for Asking Portfolio Questions (edit original post)

User avatar
Raybo
Posts: 1659
Joined: Tue Feb 20, 2007 11:02 am
Location: San Francisco
Contact:

Re: [Book: Living off Your Money, by M. McClung (Prime Harvesting)]

Post by Raybo » Thu Jul 21, 2016 3:49 pm

I contacted Michel McClung about the tests Siamond ran and graphed above and the conclusions that were drawn. Here is his response:
I believe the initial criteria for this type of analysis is to first
establish a metric. Without a metric there is no practical way to compare
MANY retirements across multiple datasets, asset allocations, and market
simulations (e.g. bootstrapping). I picked two related metrics: 1) the
maximum sustainable withdrawal rate (MSWR) with fixed withdrawals, and 2)
the success rate with fixed withdrawals. These are the same metrics used in
virtually all retirement research. Also, I believe they have a strong
correlation to variable-withdrawal results.

It is with the above metrics that Prime Harvesting is the top performer
across multiple datasets and bootstrapping simulations (i.e. resampling). So
if you don't trust these metrics then I understand you won't trust my
results, but then a new problem arises: what is a better metric?

I'm not aware of better metrics for general comparisons. MSWR and success
rate are easy to understand and generally reflect well the risk of
insufficient income...using both fixed withdrawals and variable withdrawals.
(I should make clear, I define retirement risk as the risk of insufficient
income, or more simply the risk of running out of money prematurely. There
are other risks, but I believe this is the dominant one.)

All that said, I agree with Siamond that in an ideal world all my research
would be based on an agreed-on variable withdrawal metric. I actually
define a candidate variable-withdrawal metric in the book (i.e. HREFF) to
compare variable-withdrawals strategies, so I could have easily have used
it, but I believe doing so for harvesting methods like Prime Harvesting
would hinder the reader more that helping (i.e. it's complex and hasn't
been established in the industry). I did use HREFF to evaluate Prime
Harvesting in my own research, but only to confirm the main results.

As for alternatives, results are shaped by what is measured. I don't see the
stock-bond ratio as a good metric for retirement risk (it doesn't correlate
well with insufficient income). I also don't see portfolio value as a good
metric (i.e. the goal typically isn't to maximize value, or even income, but
to minimize the risk of insufficient income).
No matter how long the hill, if you keep pedaling you'll eventually get up to the top.

User avatar
siamond
Posts: 4155
Joined: Mon May 28, 2012 5:50 am

Re: [Book: Living off Your Money, by M. McClung (Prime Harvesting)]

Post by siamond » Thu Jul 21, 2016 4:20 pm

Raybo wrote:I contacted Michel McClung about the tests Siamond ran and graphed above and the conclusions that were drawn. Here is his response: [...]
Thank you for doing that. I actually had a private discussion with the author a few weeks ago along the same lines. I do understand where he is coming from. He's right that a lot of retirement research is based on the SWR, but this doesn't change the fact that this is a pretty much meaningless metric (it's just easy to measure and understand, which is why it is widely used!). And he's certainly right that it is difficult to find a solid metric in order to do a better assessment.

This is why my table of scenario results (for those of you who clicked on it in this post) tracks numerous metrics, because I just can't find one that properly summarizes all the aspects of what I view as a good withdrawal (or harvesting) method. Even that doesn't replace the need to take a look at multiple representative cycles. This is certainly not very satisfying, and I'll definitely take a closer look at HREFF. Although I find hard to believe that a single metric can capture it all, while addressing diverse sets of goal and definitions of risk.

Actually, the more I think about it, the more it seems to me that we should start by the withdrawal methods. See if McClung's ideas (e.g. EM and so on) hold against solid methods like VPW and PMT derivatives, or Guyton-Klinger decision rules. And then and only then, ponder if Prime Harvesting adds value to those withdrawal methods. The question is not about Prime Harvesting in isolation (well, combined with a very suboptimal fixed withdrawal method to be more precise), the question is really Prime Harvesting possibly combined with a sensible withdrawal method.

EDIT: I meant 'SWR' instead of 'MSWR'. Fixed.
Last edited by siamond on Sun Jul 31, 2016 10:57 am, edited 1 time in total.

AlohaJoe
Posts: 3684
Joined: Mon Nov 26, 2007 2:00 pm
Location: Saigon, Vietnam

Re: [Book: Living off Your Money, by M. McClung (Prime Harvesting)]

Post by AlohaJoe » Thu Jul 21, 2016 9:56 pm

longinvest wrote:It would be quite interesting to see the portfolio and withdrawal trajectory of 30/70, 40/60, and maybe 50/50 (stocks/bonds) constant AA portfolios vs Prime Harvesting in both nominal and inflation-adjusted term, and look at the result with the eye of a retiree who sees nominal numbers on his bank account and mutual fund statements.
FWIW, I do usually look at nominal values but I don't usually write/post pictures of them. Partly it is an attempt to not overwhelm people with 100 different charts and metrics, which anyone (even me) probably couldn't make a lot of sense of. But I think you do have a point that there is probably too much of a reliance on inflation-adjusted figures. I mean, it isn't like I have any actual idea what the inflation-adjusted to the year 2000 value of my portfolio is. I don't look at CPI every year and adjust my portfolio. I don't look for when CPI inevitably gets revised a few times. I'm guessing that is by far the most common way, since it requires the least amount of work :)

heyyou
Posts: 3095
Joined: Tue Feb 20, 2007 4:58 pm

Re: [Book: Living off Your Money, by M. McClung (Prime Harvesting)]

Post by heyyou » Thu Jul 21, 2016 11:13 pm

McClung repeatedly suggested the use of real, guaranteed income to cover all or most of a retiree's necessary expenses. The initial bond allocation (45%) and the initial stock allocation (55%) are not your only retirement income income sources, and you have already spent some of your portfolio on buying guaranteed income. The book is about how to better harvest from your equities, AFTER you have considered how to cover your necessary expenses.

He states that starting with stock allocations higher than 55%, increases risk due to the volatility. So more stocks is counterproductive to portfolio longevity, when not rebalancing annually. The choices are more income in retirement starting with 45% in bonds, and spending from them first, and seldom rebalancing, or annual rebalancing to buffer portfolio volatility. The 2007 Spitzer and Singh paper that questioned rebalancing in retirement was thorough, and it too was data driven, not tuned by what would suit the behavioral aspects of retirees.

MSWR is based on surviving the single worst period, then posters pick that same period to show that it is not a good period for McClung's suggestions. It is not a good period for any retirement. One of McClung's points is that his methods are an improvement over MSWR, allowing the retiree to spend more from the same amount of assets, with more assurance that it will survive a similar worst period, as does VPW.

In the book, he did use beneficial practices from several historical WD methods, showing what parts he kept and what parts were inadequate. Mortality is included in Table #58 which is used to make minor adjustments to the WD amount, same as VPW. McClung's work is an improvement over the most popular and dated spending methods, as is VPW.

AlohaJoe
Posts: 3684
Joined: Mon Nov 26, 2007 2:00 pm
Location: Saigon, Vietnam

Re: [Book: Living off Your Money, by M. McClung (Prime Harvesting)]

Post by AlohaJoe » Fri Jul 22, 2016 12:15 am

heyyou wrote:McClung repeatedly suggested the use of real, guaranteed income to cover all or most of a retiree's necessary expenses.
This was not my interpretation of the book, nor of my private conversations with the author.

Page 8: "Guaranteed income still has an optional role, but it must be a balanced one."

Page 9: (in the one-page reference guide) "Optionally consider using guaranteed income"

Page 115: when speaking about variable withdrawal strategies "(It makes sense for retirees to set the income floor to their essential-income rate, but more on this later.)"

Page 192: "As a general rule, let systemic withdrawals do what it does well (managing volatile portfolio, producing strong income) by providing it with the best possible parameters to work from, then use guaranteed income to reduce risk further if desired."

Page 263: "Is it prudent for every retiree to use GI? No, not necessarily. This decision depends on the many interrelated strands that determine how GI fits and functions in retirement, one of which is a ordability."

Page 265: "Do these systemic withdrawal numbers imply that the results are good enough to eliminate the need for guaranteed income? Based on known risk, yes"

Page 265: "Although best practices don’t remove the justi cation for GI, they do remove any perceived mandate that existed for using it. A retiree can create a reasonable plan without GI. GI may be desirable at times, depending on a retiree’s circumstances and current market conditions, but the numbers show it’s optional."

Page 273: "Since GI is often not cheap, the typical goal is to purchase only enough GI to ensure that essential expenses are covered. As previously mentioned, however, the amount of GI purchased doesn’t have to match essential expenses. It’s reasonable to count on systemic withdrawals to provide some income, even during a failure market."

Page 273: "These typical costs provide a useful gauge for deciding how much, if any, GI to purchase."

Page 274: "Purchasing GI only makes sense if it’s a affordable."

Page 314: "The primary recommendations follow:
1. A. When valuations are high at the start of retirement (i.e., the valuation level is 3 or above)
a. Immediately lock in a minimum of 50% of the essential-income rate with in inflation-adjusted SPIAs."

(i.e. the retiree does not have a real, guaranteed income to cover all or most of the necessary expenses)

Page 314: "Retirees can then decide where they fit on the spectrum of deploying GI, if at all."

Page 327: "EM with no GI almost always performs better than ECM with GI...Based purely on known market risk, GI can be justified only if both the valuation level and interest rates are clearly high, but this is rare."

Pages 335-341 have several cases where the author does not recommend GI.

NMJack
Posts: 836
Joined: Sun Feb 14, 2016 1:22 pm

Re: [Book: Living off Your Money, by M. McClung (Prime Harvesting)]

Post by NMJack » Fri Jul 22, 2016 2:11 am

heyyou wrote:McClung repeatedly suggested the use of real, guaranteed income to cover all or most of a retiree's necessary expenses.
This comment reminded me of the old Steve Martin routine regarding "how to be a millionaire, and never pay taxes." Any other old people remember that one? I think the album was titled "comedy is not pretty."

Post Reply