endowment method of retirement spending

[starboardtack]

Hello all.  My first post/question on this fine forum.  Am retired early @ 60 and in distribution phase of IRA and the question relates to same circumstance.   I seek comments and observations regarding adaptation of the so called "endowment" model of spending to retirement fund spending.  The model is used for endowments at various universities, and mentioned by Robert Carlson in New Rules of Retirement, and explained to me by economics prof (aka crew member in training).   As all know, the goal of endowments is longevity of spending.  Assuming the flexibility in spending to accomodate the model:

Last year's spending + inflation X 70 % = $pot #1 
Target distribution rate X retirement fund value X 30 % = $pot #2

$pot #1 + $pot #2 = distribution for the year

This model seems to avoid the "wasting" effect of common fixed percentage + inflation spending which increases each year regardless of portfolio performance and burdens the portfolio.  Rather, it adjusts spending to reflect portfolio performance without exceptional pain in years of poor performance while allowing moderate exuberance, and thus avoiding deprivation, in good times.  As has been pointed out by the model's proponents, the retirement fund, unlike the endowment fund, does not have to last forever.  Thus, additional flexibility is afforded, when needed, through temporary suspension of the spending model if a successive decline in necessary spending is caused by several years of poor performance and some spending of principal is required.  Subsequent spending after resumption of the model would reflect this contingency since 30 % of spending is based upon the value of the fund.   The model benefits from the many decades of its use in spending of endowment funds and has withstood the scrutiny many professors of economics, some of whom have benefited from the spending.     
Any comments would be welcome.

 

[jeff2006]

A similar model is discussed at length by Henry Hebeler at www.analyzenow.com

Go to the section on "free publications" and look at the articles called "The Biggest Retirement Planning Myth" and "Financial Dynamics Improves Lifelong Planning."

Mr. Hebeler calls the model the "Retirement Autopilot."  His ideas include further refinements that take finite lifespan into account wrt the target protfolio withdrawal rate.

Despite the title of the first-cited work above, he is a great supporter of your model -- the title suggests criticism of competing techniques, not the one you have proposed.

In my opinion, Hebeler's work is highly credible.

I hope this helps . . .

 

[d]

this methods seems to be directly equivalent to the fixed/variable method, but specifying a 70/30 initial mix

 

[justin]

this methods seems to be directly equivalent to the fixed/variable method, but specifying a 70/30 initial mix

Those were my thoughts too at first, but it's a little more complicated than that. The 70% component is based on last year's spending (based on the formula) plus inflation, whereas the fixed component of the fixed+variable method is a predetermined amount set in year zero of the withdrawal strategy.

The "endowment method" seems to provide an additional dampening effect on upward and downward revisions in withdrawals since the n+1 year withdrawal is linked to the n year withdrawal. The fixed+variable method has no such year to year link (other than partial portfolio depletion due to previous year's spending).

Seems like an interesting strategy overall though.

 

[d]

ah yes!

 

[starboardtack]

Thanks for the reference and link, Jeff.  I especially like the "retirement autopilot" analogy to which Mr. Hebeler refers.  I've bumped it up on my "to read" list.

The dampening effect caused by the linkage of n+1 year withdrawal and n year withdrawal is what piqued my interest, Justin.  It seems a "kinder and gentler" method of course adjustments to the spending of a retirement fund.  I intend to adopt this model in the coming months.  Some have pointed to concerns for this approach during times of stagflation.  However, the inherent flexibility helps. 

Of course, I can always move aboard, ease the mainsheets, hoist the halyard, and cast off!     That would be not a bad thing in any event.   

 

[sailor]

Of course, I can always move aboard, ease the mainsheets, hoist the halyard, and cast off! That would be not a bad thing in any event.

Welcome to the board starboard tack.
The additional beauty of this solutuin is that for most people liveaboard expenses are smaller than for dirt dwellers
What kind of a boat?

 

[starboardtack]

Cutter rigged Shannon 38.  Harder it blows, the better she likes it.      Thinking of casting off for a year or so, first to Mexico, then a "puddle jump" to So. Pacific.  Thus the appeal of the "retirement autopilot" analogy.

 

[sailor]

beautiful boat, wish you good luck on your journey plans.
And it looks like a lot of brightwork to keep you occupied

We are a family of 3 (soon to be 4), and are planning to cast off in maybe 10-12 years. (sooner if we attain FI sooner)
probably ICW & Carribean first.
DW likes the space and speed of cats, so maybe a medium sized cruising cat is in our future,
like a Gemini 3400 or one of smaller PDQs.
Currently along with LBYM principle we satisfy our sailing lust with a trailerable pocket cruiser.

 

[Chris24]

Starb, links not working. Could you repost link ... THanks

 

[lazygood4nothinbum]

Cutter rigged Shannon 38.  Harder it blows, the better she likes it.      Thinking of casting off for a year or so, first to Mexico, then a "puddle jump" to So. Pacific.  Thus the appeal of the "retirement autopilot" analogy.

i chatted with walt schulz aboard his highly stylized 38 ft motor boat at fort laud boat show a few years back. interesting guy. he is all boat and looks the part. very proud of his boats & with good reason. i haven't been on a shannon sail but his motor boat was gorgeous.

i'm thinking of moving aboard within 5 years and i'd also like my investments to be as automated as possible so appreciate you bringing up this topic. good info.

 

[starboardtack]

beautiful boat, wish you good luck on your journey plans.
And it looks like a lot of brightwork to keep you occupied 

Currently along with LBYM principle we satisfy our sailing lust with a trailerable pocket cruiser.

Thanks Sailor.  Got eager young crew for the bright work.    Sailed cats in the Antilles and around Barbados.  Great fun but prefer a keel under me arse in a big blow.  Brother lives on ICW in Wilmington, NC. and the shifting sandbars there make things exciting.  Got to make the passage soon, swallow the anchor, or trade for a stinkpot.  Not old enought for that yet.   Fair winds.

Starb, links not working. Could you repost link ... THanks

Chris the link I referred to was  Henry Hebeler's at www.analyzenow.com posted by Jeff in this this thread.  The book I referred to was "The New Rules of Retirement; Strategies for a Secure Future" by Robert Carlson, JD, CPA.  Hope these are the links/references you wanted.

...he is all boat and looks the part. very proud of his boats & with good reason. i haven't been on a shannon sail but his motor boat was gorgeous.

i'm thinking of moving aboard within 5 years and i'd also like my investments to be as automated as possible so appreciate you bringing up this topic. good info.

You are welcome.  Shannon boats are robust and put together well, IMHO.  Wouldn't  call it a light airs boat, but a very suitable, stable and safe bluewater world cruiser.  A well behaved lady, especially in heavy weather.  ERed from high tech, high stress place and wanted to simplify.  Didn't want to constantly tinker with std. dev., SWRs, etc.  Did a diverse Vanguard based allocation similar to what Bernstein outlined in "The Four Pillars of Investing".  Living costs decrease very nicely if living aboard.  Can live fairly high on the hog (around the ears somewhere) when not a dirt dweller.    Will be moving back aboard in a couple of months after landlines are stowed.  Regards.

 

[bosco]

I think we have some sort of a record here in thread-hijacking. Not sure how to measure it, but the hijack was swift and total....

 

[bosco]

or does it officially qualify as a hijack when the op participates?

 

[cute fuzzy bunny]

It does, but there are far more impressive examples of hijacks. A few surprised even me, and some of them were my posts, and on a couple I didnt even do the hijacking!

 

[starboardtack]

I think we have some sort of a record here in thread-hijacking.  Not sure how to measure it, but the hijack was swift and total....

It does, but there are far more impressive examples of hijacks. A few surprised even me, and some of them were my posts, and on a couple I didnt even do the hijacking!

Uh Oh!  Busted by the thread police?   : What rule(s) did I violate?   Just answered some unsolicited queries as to "retirement autopilot" from those who contemplate not sitting in front of a monitor for the duration of their retirement.   

 

[dory36]

Ignore the landlubbers. They just don't understand that cruising is a suitable diversion in any thread...

dory36

 

[starboardtack]

Ignore the landlubbers. They just don't understand that cruising is a suitable diversion in any thread... 

dory36

Roger that Dory, thanks.  Been messing with the fickle iron jib all day.     Which is why I might be a little cranky.  May just use it as a sea anchor.  Will mend my ways.     After an almost last qtr. moonlight sail.    

 

[obryanjf]

Those were my thoughts too at first, but it's a little more complicated than that. The 70% component is based on last year's spending (based on the formula) plus inflation, whereas the fixed component of the fixed+variable method is a predetermined amount set in year zero of the withdrawal strategy.

The spending from last year ='s last years withdrawl, at least in my book.  Then the Hybrib and Endowment are the same with 70/30 versus my typical 50/50.

The "endowment method" seems to provide an additional dampening effect on upward and downward revisions in withdrawals since the n+1 year withdrawal is linked to the n year withdrawal. The fixed+variable method has no such year to year link (other than partial portfolio depletion due to previous year's spending).

I'm not sure I agree with this.  The fixed portion is linked to last year as it forms the starting point then gets inflation adjsuted.  The way I think of it is part is inflation adjusted and part is portfolio adjusted.

The way I read above and my code makes it equivalent to Hybrid with different percents.  Perhaps I've coded it wrong..would not be the 1st time.

I don't think this is the same as Hebener.  If been meaning to compare his but have not messed with it yet.

job

 

[justin]

Job,

Maybe you and I have a different conception of what the "hybrid method" is.

I assume it is something like taking a 3% "fixed" withdrawal each year based on the original principal amount (in year 0). Each year it is adjusted for inflation. The "fixed" component of the withdrawal does not change based on fluctuations in portfolio value from year to year.

The "variable" component of the withdrawal is, say, 1.5% of the portfolio value each year. Regardless of inflation or portfolio value fluctuation.

The real value of the fixed component of the hybrid method of withdrawal is established in year zero for all subsequent years. In nominal terms, the only thing that changes the fixed component is the CPI inflator value for each future year.

In contrast, the endowment method defines the "fixed" component of withdrawal each year in terms of the previous year's withdrawal (which includes the "fixed" component and the "variable" component). In other words, the "fixed" component includes feedback from the portfolio value fluctuation.

Under my version of the hybrid method, the "fixed" component is not affected by the fluctuation in portfolio value.

 

[OldAgePensioner]

Jeff2006,
thanks for that link those are very thought provoking articles.

 

[obryanjf]

Justin, you are right, The hybrid has a fixed and variable component.  The fixed is set at the start and is inflation adjusted.  The var is reset every year based on the ending portfolio value.  I have coded the endowement as fixed based on last years draw * percentage + variable component * percentage. 

They act a little different.  When i turn on a failure at a draw of < 50% of initial is on, the endowement performs very well (see graph).  When that is turned on then they perform about the same. The average of the terminal value and the std dev of NW is very similar.  Judge for yourself though.

job

BTW these runs have no extras and no Ty and no parttime.

< 50% draw is a failure


<50% draw is not a failure

 

[justin]

So the dampening effect of the endowment method allows a higher withdrawal rate when "failure" is defined as complete portfolio depletion.

 

[obryanjf]

Yes, that is correct.  Just don't forget with the defination of failure being running out of money, withdrawls going down to $1 and bouncing back up, is not failure, but I'm sure it would not feel good.  That is why I added the < 50% rule to count as a failure. 

I do not quite understand the WD% reaction and will need to think about it a bit.  It does a good job of keeping the WD% in a range which I think is important.  See the graph.  For comps I'll add a graph of the Hybrid 50/50. 

WD = withdrawal; 1MM port, 6% initial draw, Non monte carlo mode, no failure on < 50%, 40 year simulation

4 graphs, Hybrid WD%'s over time and WD's over time and the same for the Endowement(EM) so 4 graphs total.

The EM adjusted down further than the hybrid, notice on the WD graphs the lower amounts during the middle of the sim.  The WD% on the EM stays in a nive range.  This is similar behaviour as the ESRbob model. 

Comments

job

 

[justin]

It appears image 1 and 3 are the same, and 2 and 4 are the same.

From the endowment method graphs, it looks like the withdrawal rate will be 5-7% most years. That's 5-7% of the initial portfolio balance?