Safemax with varying AA

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
Post Reply
gilgamesh
Posts: 835
Joined: Sun Jan 10, 2016 9:29 am

Safemax with varying AA

Post by gilgamesh » Mon Jun 12, 2017 10:13 am

Is there any study that crunched the SAFEMAX (SWR) when AA changed over time...say for the first 10 years 40% stocks, next 10 with 50% and the last 10 with 60% stocks. Or variations of those, like 40% stocks for the first 10 years and then 60% stocks for 20 years.

I know of several studies had for various asset allocation but was fixed throughout. I also know of Kitces article on dynamic asset allocation. But, I am looking for real life situation where retirees go bond heavy closer to retirement when sequence risk is highest.

Thanks!

gilgamesh
Posts: 835
Joined: Sun Jan 10, 2016 9:29 am

Re: Safemax with varying AA

Post by gilgamesh » Mon Jun 12, 2017 8:08 pm

Is this a useless metric? Is no one interested in such an analysis?

It is known retirees are most vulnerable to sequence risk during the first 10 years of retirement. One way to hedge this is to go bond heavy during that period. So, isn't it critical to know how past rolling 30 year periods would have fared, if a retiree reduced their stock exposure to 30% or 40% for the first 10 years. I am surprised why there has been no such analysis.

Once the above is determined, I was curious to learn how hedging inflation risk would have affected SAFEMAX. Say by having 50% or even 100% of the bond allocation in TIPS. I know TIPS is relatively new, but if there were any other indicators of governments future inflation prediction, then comparing that to nominal treasuries should give the expected TIPS yield. I know out of control inflation affected SAFEMAX significantly during certain 30 year rolling periods.

The above two will hedge against a stock collapse and high inflation for the ten years and IMO should increase the SAFEMAX to be above 4%. If not, at least isn't it worth studying this?

I am curious what others think?

Thanks!

NibbanaBanana
Posts: 194
Joined: Sun Jan 22, 2017 10:34 pm

Re: Safemax with varying AA

Post by NibbanaBanana » Mon Jun 12, 2017 9:29 pm

Interesting. I don't know anything about this but I would think that the reason you would be able to increase your withdraw rate would be that your portfolio has grown over time due to stocks. Your idea would avoid having to sell stocks in a market collapse but there won't be as many stocks to recover and make gains. And bonds are only paying 240 bp now. So, I don't know.

bigred77
Posts: 1890
Joined: Sat Jun 11, 2011 4:53 pm

Re: Safemax with varying AA

Post by bigred77 » Mon Jun 12, 2017 9:51 pm

There have been a couple of projects that I am aware of that look into a rising equity glidepath in retirement. I've read some of the material but didn't find the conclusions all that convincing (it's been a while though).

I don't know if your parameters were specifically used but you can google "rising equity glidepath in retirement" and read through some of the material.

AlohaJoe
Posts: 2423
Joined: Mon Nov 26, 2007 2:00 pm
Location: Saigon, Vietnam

Re: Safemax with varying AA

Post by AlohaJoe » Mon Jun 12, 2017 10:05 pm

gilgamesh wrote:Is there any study that crunched the SAFEMAX (SWR) when AA changed over time...say for the first 10 years 40% stocks, next 10 with 50% and the last 10 with 60% stocks. Or variations of those, like 40% stocks for the first 10 years and then 60% stocks for 20 years.
There are a million variants on "AA changes over time" so I'm not sure anyone has looked at something exactly like what you describe.

But you could start by looking at

"Revisiting the Optimal Distribution Glidepath", David Blanchett
"Initial Conditions and Optimal Retirement Glidepaths", David Blanchett
"Dynamic Allocation Strategies for Distribution Portfolios", David Blanchett
"Confirming the Value of Rising Equity Glide Paths: Evidence from a Utility Model", Luke Delorme

itstoomuch
Posts: 4554
Joined: Mon Dec 15, 2014 12:17 pm
Location: midValley OR

Re: Safemax with varying AA

Post by itstoomuch » Tue Jun 13, 2017 12:22 am

gilgamesh wrote:Is there any study that crunched the SAFEMAX (SWR) when AA changed over time...say for the first 10 years 40% stocks, next 10 with 50% and the last 10 with 60% stocks. Or variations of those, like 40% stocks for the first 10 years and then 60% stocks for 20 years.

I know of several studies had for various asset allocation but was fixed throughout. I also know of Kitces article on dynamic asset allocation. But, I am looking for real life situation where retirees go bond heavy closer to retirement when sequence risk is highest.

Thanks!
If some one has developed a failsafe SWR plan, they oughta patent it and sell it through financial advisors :mrgreen: :idea:
Rev90517; 4 Incm stream buckets: SS+pension; dfr'd GLWB VA & FI anntys, by time & $$ laddered; Discretionary; Rentals. LTCi. Own, not asset. Tax 25%. Early SS. FundRatio (FR) >1.1 67/70yo

gilgamesh
Posts: 835
Joined: Sun Jan 10, 2016 9:29 am

Re: Safemax with varying AA

Post by gilgamesh » Tue Jun 13, 2017 10:07 am

NibbanaBanana wrote:Interesting. I don't know anything about this but I would think that the reason you would be able to increase your withdraw rate would be that your portfolio has grown over time due to stocks. Your idea would avoid having to sell stocks in a market collapse but there won't be as many stocks to recover and make gains. And bonds are only paying 240 bp now. So, I don't know.
In the current environment bond yields are low, however stocks are widely considered to be at peak too given P/E ratio, so it's a mixed bag. However, SAFEMAX is independent of any of these starting parameters, which is the most critical point. Advocates of SAFEMAX claim, as it was the worst performance in history, irrelevant of your current starting point, it will most likely hold up in the future.

Increasing bond allocation close to retirement is not my idea, it's widely known.

gilgamesh
Posts: 835
Joined: Sun Jan 10, 2016 9:29 am

Re: Safemax with varying AA

Post by gilgamesh » Tue Jun 13, 2017 10:15 am

bigred77 wrote:There have been a couple of projects that I am aware of that look into a rising equity glidepath in retirement. I've read some of the material but didn't find the conclusions all that convincing (it's been a while though).

I don't know if your parameters were specifically used but you can google "rising equity glidepath in retirement" and read through some of the material.
Yes, Kitces and Pfau came up with the rising equity glidepath. Yes, AFAIK Kitces larer concluded there was nothing special about the rising equity glidepath, except the expected increase in risk/return with increased exposure to stocks.

As far as I am concerned rising equity glidepath has no significance.

dbr
Posts: 23323
Joined: Sun Mar 04, 2007 9:50 am

Re: Safemax with varying AA

Post by dbr » Tue Jun 13, 2017 10:23 am

I suppose that like a principle of least action one could imagine writing formulas for a path in time through allocation space, more likely a space of expected return and expected risk, that optimizes some quantity. I am not aware of anyone doing that and the idea remains hazy as to what would actually be constructed.

gilgamesh
Posts: 835
Joined: Sun Jan 10, 2016 9:29 am

Re: Safemax with varying AA

Post by gilgamesh » Tue Jun 13, 2017 10:28 am

AlohaJoe wrote:
gilgamesh wrote:Is there any study that crunched the SAFEMAX (SWR) when AA changed over time...say for the first 10 years 40% stocks, next 10 with 50% and the last 10 with 60% stocks. Or variations of those, like 40% stocks for the first 10 years and then 60% stocks for 20 years.
There are a million variants on "AA changes over time" so I'm not sure anyone has looked at something exactly like what you describe.

But you could start by looking at

"Revisiting the Optimal Distribution Glidepath", David Blanchett
"Initial Conditions and Optimal Retirement Glidepaths", David Blanchett
"Dynamic Allocation Strategies for Distribution Portfolios", David Blanchett
"Confirming the Value of Rising Equity Glide Paths: Evidence from a Utility Model", Luke Delorme
I am not talking about the millions of variations, just two options.

Keep asset allocation the same from day of retirement to death ....This is the only SAFEMAX data we have. The second option when a retiree reduces their stock exposure when they near retirement and afterwards. I think this is the more common approach and yet no SAFEMAX data on it.

I am not a fan of rising equity glidepath, I prefer the more recent Kitces concept of the bond tent

https://www.kitces.com/blog/managing-po ... -red-zone/

I am not interested in dynamic asset allocation based on P/E ratio and optimizing SAFEMAX which has also been studied and published by Kitces.
Last edited by gilgamesh on Tue Jun 13, 2017 10:45 am, edited 1 time in total.

gilgamesh
Posts: 835
Joined: Sun Jan 10, 2016 9:29 am

Re: Safemax with varying AA

Post by gilgamesh » Tue Jun 13, 2017 10:34 am

dbr wrote:I suppose that like a principle of least action one could imagine writing formulas for a path in time through allocation space, more likely a space of expected return and expected risk, that optimizes some quantity. I am not aware of anyone doing that and the idea remains hazy as to what would actually be constructed.
Yes! That would be optimal. But I'm not asking for that.

Just any SAFEMAX data that does not keep AA constant throughout retirement. I would think keeping AA constant through the retirement would be more rare than having it more conservative immediately after retirement, no?

P.S: What you have described above is the gold standard....the perfect look back bias...SAFEMAX is nothing but look back bias, and what you've described will perfect it. I'm not looking for perfection, but something that does not fix AA throughout retirement.

gilgamesh
Posts: 835
Joined: Sun Jan 10, 2016 9:29 am

Re: Safemax with varying AA

Post by gilgamesh » Tue Jun 13, 2017 10:40 am

itstoomuch wrote:
gilgamesh wrote:Is there any study that crunched the SAFEMAX (SWR) when AA changed over time...say for the first 10 years 40% stocks, next 10 with 50% and the last 10 with 60% stocks. Or variations of those, like 40% stocks for the first 10 years and then 60% stocks for 20 years.

I know of several studies had for various asset allocation but was fixed throughout. I also know of Kitces article on dynamic asset allocation. But, I am looking for real life situation where retirees go bond heavy closer to retirement when sequence risk is highest.

Thanks!
If some one has developed a failsafe SWR plan, they oughta patent it and sell it through financial advisors :mrgreen: :idea:
One could say a 3% SWR is for all intense and purposes the failsafe SAFEMAX, which theoretically allows withdrawal to perpetuity...or is it 3.5%. I am not necessarily looking for the perfect solution. But, I am looking for parameters that would increase SWR, as per Bengen study criteria.

I know sequence risk is highest 10 years after retirement. I am also fairly certain retirees keep a more conservative AA around retirement (The bond tent). Therefore, wouldn't a SAFEMAX calculated based out of this be more representative than all the SAFEMAX data out there which keeps AA the same through retirement.

P.S: I think the 4% SWR is based on something like 60% stocks 40% bonds (or some constant AA), will the 4% SWR be higher if it was more conservative for the first 10 years of retirement and had some inflation hedging in the form of TIPS... Isn't that a reasonable question to ask...no, an important question to ask?

itstoomuch
Posts: 4554
Joined: Mon Dec 15, 2014 12:17 pm
Location: midValley OR

Re: Safemax with varying AA

Post by itstoomuch » Tue Jun 13, 2017 11:22 am

The only method that I have found, but not exploring, was to continually & periodically check on Funding Ratio. Anything above 1.1 get reallocated to another bucket. We have been lucky to reallocate just once, 2008, the other direction into real life annuities to assure a minimum FR.

Thanks to a long bull market, our Standard of living lifefestyle has increased because the asset base has been increasing in the annuity guaranteed stepups & market roll-ups.

Ymmv
Rev90517; 4 Incm stream buckets: SS+pension; dfr'd GLWB VA & FI anntys, by time & $$ laddered; Discretionary; Rentals. LTCi. Own, not asset. Tax 25%. Early SS. FundRatio (FR) >1.1 67/70yo

itstoomuch
Posts: 4554
Joined: Mon Dec 15, 2014 12:17 pm
Location: midValley OR

Re: Safemax with varying AA

Post by itstoomuch » Tue Jun 13, 2017 11:27 am

When I started to think about this, 2007-2008, as I was caring for my mother and in-laws, I looked at "option" diagrams to see what would happen to our asset value in various markets. :) .
Ymmv
Rev90517; 4 Incm stream buckets: SS+pension; dfr'd GLWB VA & FI anntys, by time & $$ laddered; Discretionary; Rentals. LTCi. Own, not asset. Tax 25%. Early SS. FundRatio (FR) >1.1 67/70yo

AlohaJoe
Posts: 2423
Joined: Mon Nov 26, 2007 2:00 pm
Location: Saigon, Vietnam

Re: Safemax with varying AA

Post by AlohaJoe » Tue Jun 13, 2017 11:32 am

gilgamesh wrote:I think this is the more common approach and yet no SAFEMAX data on it.
Calculating SAFEMAX isn't hard so here it is, since you seem keen to have it. Using Shiller's annual data, the rising equity glidepath you describe has a SAFEMAX of 3.70%. A 60/40 portfolio has a SAFEMAX of 3.66%.

itstoomuch
Posts: 4554
Joined: Mon Dec 15, 2014 12:17 pm
Location: midValley OR

Re: Safemax with varying AA

Post by itstoomuch » Tue Jun 13, 2017 11:34 am

Finally, once one achieves a secure FR, one can do any thing with the excess assets. For me, I acquired more equities in Discretionay acct (rising equity portfolio in retirement, see threads on this). :greedy
Rev90517; 4 Incm stream buckets: SS+pension; dfr'd GLWB VA & FI anntys, by time & $$ laddered; Discretionary; Rentals. LTCi. Own, not asset. Tax 25%. Early SS. FundRatio (FR) >1.1 67/70yo

gilgamesh
Posts: 835
Joined: Sun Jan 10, 2016 9:29 am

Re: Safemax with varying AA

Post by gilgamesh » Tue Jun 13, 2017 12:35 pm

AlohaJoe wrote:
gilgamesh wrote:I think this is the more common approach and yet no SAFEMAX data on it.
Calculating SAFEMAX isn't hard so here it is, since you seem keen to have it. Using Shiller's annual data, the rising equity glidepath you describe has a SAFEMAX of 3.70%. A 60/40 portfolio has a SAFEMAX of 3.66%.
Thank you. Is the 3.7% for 40/60 for the first 10 years and then 60/40 for the next 20?

I assume you did it for what you thought would be the worst 30 year period. If so, how did you know which year will show the worst performance for 'rising equity glidepath'? Doesn't this vary based on AA? Or did you do all 30 year periods from 1926-2010 or some similar range?

gilgamesh
Posts: 835
Joined: Sun Jan 10, 2016 9:29 am

Re: Safemax with varying AA

Post by gilgamesh » Tue Jun 13, 2017 12:38 pm

itstoomuch wrote:Finally, once one achieves a secure FR, one can do any thing with the excess assets. For me, I acquired more equities in Discretionay acct (rising equity portfolio in retirement, see threads on this). :greedy
I of course have a similar plan, and I'm trying to see how it fits with the traditional SAFEMAX SWR concept.

Will flooring with TIPS from retirement age of 60 to SS at age 70 (happens to be the first 10 years of retirement) also result in higher SWR, as viewed through the traditional SWR withdrawal.

MIpreRetirey
Posts: 632
Joined: Fri Sep 06, 2013 12:35 pm

Re: Safemax with varying AA

Post by MIpreRetirey » Tue Jun 13, 2017 1:01 pm

gilgamesh wrote:
itstoomuch wrote:Finally, once one achieves a secure FR, one can do any thing with the excess assets. For me, I acquired more equities in Discretionay acct (rising equity portfolio in retirement, see threads on this). :greedy
I of course have a similar plan, and I'm trying to see how it fits with the traditional SAFEMAX SWR concept.

Will flooring with TIPS from retirement age of 60 to SS at age 70 (happens to be the first 10 years of retirement) also result in higher SWR, as viewed through the traditional SWR withdrawal.
I'd look at the flooring is still part of your total portfolio as far as total WR. But, you can consider it a source of inflation adjusted income (after subtracting it's initial total cost from your total portfolio) in firecalc.com. On the last tab in firecalc, you can investigate the best AA for your WR and time needed, and other things.
I often run two firecalcs: one for pre-SS, and then one for a new WR after SS.
new username.

AlohaJoe
Posts: 2423
Joined: Mon Nov 26, 2007 2:00 pm
Location: Saigon, Vietnam

Re: Safemax with varying AA

Post by AlohaJoe » Tue Jun 13, 2017 9:59 pm

gilgamesh wrote:Thank you. Is the 3.7% for 40/60 for the first 10 years and then 60/40 for the next 20?
No, you said "say for the first 10 years 40% stocks, next 10 with 50% and the last 10 with 60% stocks" so that's what it is for. None of the small tweaks make any difference at all.
how did you know which year will show the worst performance for 'rising equity glidepath'?
"Everyone" knows that 1966 was the worst year. Static allocation, rising glidepath, falling glidepath ... none of that changes it.
Doesn't this vary based on AA?
Nope. Well, not until you start getting below ~25% stocks, then you have so many bonds that the bond bear markets become more important.
Or did you do all 30 year periods from 1926-2010 or some similar range?
Yes, this is what I did. I used Shiller's annual data which covers 1871-2016.

itstoomuch
Posts: 4554
Joined: Mon Dec 15, 2014 12:17 pm
Location: midValley OR

Re: Safemax with varying AA

Post by itstoomuch » Tue Jun 13, 2017 10:44 pm

itstoomuch wrote:When I started to think about this, 2007-2008, as I was caring for my mother and in-laws, I looked at "option" diagrams to see what would happen to our asset value in various markets. :) .
Ymmv
I only considered equity and Not bonds. I didn't trust bonds because many GSE, MBS, SL bonds had problems and the continued willingness of Congress to fund government and debt.
YMMV.

Too many moving parts cloud the question and obsure the interpretation of any results. That's why it's called Economics.
Ymmv
Rev90517; 4 Incm stream buckets: SS+pension; dfr'd GLWB VA & FI anntys, by time & $$ laddered; Discretionary; Rentals. LTCi. Own, not asset. Tax 25%. Early SS. FundRatio (FR) >1.1 67/70yo

gilgamesh
Posts: 835
Joined: Sun Jan 10, 2016 9:29 am

Re: Safemax with varying AA

Post by gilgamesh » Wed Jun 14, 2017 8:53 am

AlohaJoe wrote:
gilgamesh wrote:Thank you. Is the 3.7% for 40/60 for the first 10 years and then 60/40 for the next 20?
No, you said "say for the first 10 years 40% stocks, next 10 with 50% and the last 10 with 60% stocks" so that's what it is for. None of the small tweaks make any difference at all.
how did you know which year will show the worst performance for 'rising equity glidepath'?
"Everyone" knows that 1966 was the worst year. Static allocation, rising glidepath, falling glidepath ... none of that changes it.
Doesn't this vary based on AA?
Nope. Well, not until you start getting below ~25% stocks, then you have so many bonds that the bond bear markets become more important.
Or did you do all 30 year periods from 1926-2010 or some similar range?
Yes, this is what I did. I used Shiller's annual data which covers 1871-2016.
Thank you so much...yes it certainly appears the glidepath had no significant effect on SWR. You are absolutely correct also in that 25% stocks and 0% stocks that had lower SWR on another year. I remembered that graph, but didn't make that relationship.

dbr
Posts: 23323
Joined: Sun Mar 04, 2007 9:50 am

Re: Safemax with varying AA

Post by dbr » Wed Jun 14, 2017 8:59 am

gilgamesh wrote:
AlohaJoe wrote:
gilgamesh wrote:Thank you. Is the 3.7% for 40/60 for the first 10 years and then 60/40 for the next 20?
No, you said "say for the first 10 years 40% stocks, next 10 with 50% and the last 10 with 60% stocks" so that's what it is for. None of the small tweaks make any difference at all.
how did you know which year will show the worst performance for 'rising equity glidepath'?
"Everyone" knows that 1966 was the worst year. Static allocation, rising glidepath, falling glidepath ... none of that changes it.
Doesn't this vary based on AA?
Nope. Well, not until you start getting below ~25% stocks, then you have so many bonds that the bond bear markets become more important.
Or did you do all 30 year periods from 1926-2010 or some similar range?
Yes, this is what I did. I used Shiller's annual data which covers 1871-2016.
Thank you so much...yes it certainly appears the glidepath had no significant effect on SWR. You are absolutely correct also in that 25% stocks and 0% stocks that had lower SWR on another year. I remembered that graph, but didn't make that relationship.

Reading through numerous articles on SWR, I know high inflation played a major role in certain 30 year periods, do you know whether it made and significant difference to the SWR for 1966?
That might be a tough question to answer. Inflation had a huge excursion between 1972 and 1982. At the same time real interest rates also had a huge upward excursion. I don't have an inflation indexed growth chart with dividends for stocks, but the period 1966-1982 was a bear market that would have done the investor no good even without a market crash. 1966 is probably the perfect storm for a bad retirement while 1982 is the perfect storm for a good retirement.

gilgamesh
Posts: 835
Joined: Sun Jan 10, 2016 9:29 am

Re: Safemax with varying AA

Post by gilgamesh » Wed Jun 14, 2017 9:06 am

MIpreRetirey wrote:
gilgamesh wrote:
itstoomuch wrote:Finally, once one achieves a secure FR, one can do any thing with the excess assets. For me, I acquired more equities in Discretionay acct (rising equity portfolio in retirement, see threads on this). :greedy
I of course have a similar plan, and I'm trying to see how it fits with the traditional SAFEMAX SWR concept.

Will flooring with TIPS from retirement age of 60 to SS at age 70 (happens to be the first 10 years of retirement) also result in higher SWR, as viewed through the traditional SWR withdrawal.
I'd look at the flooring is still part of your total portfolio as far as total WR. But, you can consider it a source of inflation adjusted income (after subtracting it's initial total cost from your total portfolio) in firecalc.com. On the last tab in firecalc, you can investigate the best AA for your WR and time needed, and other things.
I often run two firecalcs: one for pre-SS, and then one for a new WR after SS.
You know what, under the 'Other income/spending' tab, I could use the CPI adjusted pension for the TIPS floor. It will however ignore the yield which I can approximate elsewhere too. Didn't realize firecalc is this robust.

But, I still need to figure out how to mimic TIPS that vary over years to that of a pension. TIPS will be certain amount for the first ten years, then lower for the next 20 with SS and then will be zero to be replaced by an SPIA. If anyone has any tips, that'll be great, but otherwise I should be able to figure it out by trial and error. The instructions for Firecalc pension data is a little murky to me.
Last edited by gilgamesh on Wed Jun 14, 2017 9:23 am, edited 1 time in total.

gilgamesh
Posts: 835
Joined: Sun Jan 10, 2016 9:29 am

Re: Safemax with varying AA

Post by gilgamesh » Wed Jun 14, 2017 9:11 am

dbr wrote:
gilgamesh wrote:
AlohaJoe wrote:
gilgamesh wrote:Thank you. Is the 3.7% for 40/60 for the first 10 years and then 60/40 for the next 20?
No, you said "say for the first 10 years 40% stocks, next 10 with 50% and the last 10 with 60% stocks" so that's what it is for. None of the small tweaks make any difference at all.
how did you know which year will show the worst performance for 'rising equity glidepath'?
"Everyone" knows that 1966 was the worst year. Static allocation, rising glidepath, falling glidepath ... none of that changes it.
Doesn't this vary based on AA?
Nope. Well, not until you start getting below ~25% stocks, then you have so many bonds that the bond bear markets become more important.
Or did you do all 30 year periods from 1926-2010 or some similar range?
Yes, this is what I did. I used Shiller's annual data which covers 1871-2016.
Thank you so much...yes it certainly appears the glidepath had no significant effect on SWR. You are absolutely correct also in that 25% stocks and 0% stocks that had lower SWR on another year. I remembered that graph, but didn't make that relationship.

Reading through numerous articles on SWR, I know high inflation played a major role in certain 30 year periods, do you know whether it made and significant difference to the SWR for 1966?
That might be a tough question to answer. Inflation had a huge excursion between 1972 and 1982. At the same time real interest rates also had a huge upward excursion. I don't have an inflation indexed growth chart with dividends for stocks, but the period 1966-1982 was a bear market that would have done the investor no good even without a market crash. 1966 is probably the perfect storm for a bad retirement while 1982 is the perfect storm for a good retirement.
I deleted my question, you were too quick :D

After posting it I noticed that spike in 72, but I am only looking at a 10 year period after retirement where I assume the sequence risk to be high. So, for a 1966 retiree inflation will not have played a big role for the 6 years, I wonder whether the next 4 mattered much. If it was high inflation for most of the 10 years then it would have mattered more.

dbr
Posts: 23323
Joined: Sun Mar 04, 2007 9:50 am

Re: Safemax with varying AA

Post by dbr » Wed Jun 14, 2017 9:18 am

gilgamesh wrote:
I deleted my question, you were too quick :D

After posting it I noticed that spike in 72, but I am only looking at a 10 year period after retirement where I assume the sequence risk to be high. So, for a 1966 retiree inflation will not have played a big role for the 6 years, I wonder whether the next 4 mattered much. If it was high inflation for most of the 10 years then it would have mattered more.
It might be a mistake to think retirements are determined in the first ten years and that sequence of returns is the be-all and end-all risk for retirement. There are a lot of ways for things to go badly and a lot of ways for things to go well.

MIpreRetirey
Posts: 632
Joined: Fri Sep 06, 2013 12:35 pm

Re: Safemax with varying AA

Post by MIpreRetirey » Wed Jun 14, 2017 9:33 am

gilgamesh wrote:
MIpreRetirey wrote:
gilgamesh wrote:
itstoomuch wrote:Finally, once one achieves a secure FR, one can do any thing with the excess assets. For me, I acquired more equities in Discretionay acct (rising equity portfolio in retirement, see threads on this). :greedy
I of course have a similar plan, and I'm trying to see how it fits with the traditional SAFEMAX SWR concept.

Will flooring with TIPS from retirement age of 60 to SS at age 70 (happens to be the first 10 years of retirement) also result in higher SWR, as viewed through the traditional SWR withdrawal.
I'd look at the flooring is still part of your total portfolio as far as total WR. But, you can consider it a source of inflation adjusted income (after subtracting it's initial total cost from your total portfolio) in firecalc.com. On the last tab in firecalc, you can investigate the best AA for your WR and time needed, and other things.
I often run two firecalcs: one for pre-SS, and then one for a new WR after SS.
You know what, under the 'Other income/spending' tab, I could use the CPI adjusted pension for the TIPS floor. It will however ignore the yield which I can approximate elsewhere too. Didn't realize firecalc is this robust.

But, I still need to figure out how to mimic TIPS that vary over years to that of a pension. TIPS will be certain amount for the first ten years, then lower for the next 20 with SS and then will be zero to be replaced by an SPIA. If anyone has any tips, that'll be great, but otherwise I should be able to figure it out by trial and error. The instructions for pension is a little murky to me.
I think you won't be able to model TIPS very well for more than 20 years backlooking. See these two threads for efforts on TIPS: (search "TIPS" ?)
viewtopic.php?t=120430
viewtopic.php?t=2520
new username.

gilgamesh
Posts: 835
Joined: Sun Jan 10, 2016 9:29 am

Re: Safemax with varying AA

Post by gilgamesh » Wed Jun 14, 2017 9:56 am

dbr wrote:
gilgamesh wrote:
I deleted my question, you were too quick :D

After posting it I noticed that spike in 72, but I am only looking at a 10 year period after retirement where I assume the sequence risk to be high. So, for a 1966 retiree inflation will not have played a big role for the 6 years, I wonder whether the next 4 mattered much. If it was high inflation for most of the 10 years then it would have mattered more.
It might be a mistake to think retirements are determined in the first ten years and that sequence of returns is the be-all and end-all risk for retirement. There are a lot of ways for things to go badly and a lot of ways for things to go well.
That is true. However, if I believe over the long run stocks will deliver an average return of a certain amount. Say, 4% or 5% real, (The precise amount doesn't matter) on average over 30-40 years. Even if I am 100% confident on that number, it still doesn't matter in retirement due to sequence of return risk. We all assume such an average percentage during accumulation, and that average works fine as there's no sequence risk in the accumulation phase.

furthermore, Kitces for sure had concluded this sequence risk is highest for the first 10 years.

I also believe, we are sacrificing a lot of the early spending power, in being ultra conservative with our withdrawal rate, in fear of what could happen in the first 10 years in terms of sequence risk, only to have a much larger nest egg after the 10 years. However, a typical retiree would want it the other way around.

I was wondering whether there were ways to even out this spending.

IMO, the only way to do that is to hedge the sequence of return risk for the first 10 years. Will setting up an inflation adjusted floor for the first ten years accomplish this and give me a more even spending? However, this will allocate a large portion of the portfolio to bonds. So, I wanted to plug in this floor and see how it works within the traditional SWR ...it looks like Firecalc could do that. If even the traditional SWR calculations shows it to be better, then all the better.

Yes! There are other risks....but I am wanting to rely on the average stock return even in retirement.

dbr
Posts: 23323
Joined: Sun Mar 04, 2007 9:50 am

Re: Safemax with varying AA

Post by dbr » Wed Jun 14, 2017 10:05 am

I thought VPW was the so far credible answer to managing variability of returns. I haven't heard of anything better: https://www.bogleheads.org/wiki/Variabl ... withdrawal

The classic financial solution is to annuitize everything. People seem to not like that very much.

As far as sequence of returns, there are two uncertainties. One is the sequence of a known average of returns. The other is uncertainty as to the unknown future average. RTM involves the question of return to what mean. Aside from that, the data one assumes has to be real returns and/or one has to also have a sequence of inflation in mind. If sources of income and expenses mix real and nominal instruments, estimating sequence of inflation cannot be avoided.

On top of this actual retirements don't have fixed needs for income. A retiree has to estimate the variability and sequence of expenses. I think overengineering retirement is a siren song.

gilgamesh
Posts: 835
Joined: Sun Jan 10, 2016 9:29 am

Re: Safemax with varying AA

Post by gilgamesh » Wed Jun 14, 2017 11:57 am

dbr wrote:I thought VPW was the so far credible answer to managing variability of returns. I haven't heard of anything better: https://www.bogleheads.org/wiki/Variabl ... withdrawal

This is what the article says..."VPW is best used in conjunction with guaranteed base income from Social Security, pensions, and, if necessary, inflation-indexed Single Premium Immediate Annuity (SPIA).
Base income for years between retirement and the start of Social Security payments can be provided by using a simple CD ladder. For the purposes of VPW calculations, the money set aside in the CD ladder should not be considered as part of the portfolio."

VPW reacts to current market conditions, it's not free of current market conditions. Income floor is free of current and future market condition. Unlike the above suggestion of a CD ladder, a TIPS ladder could be a better choice. A TIPS floor, allows for 80% stocks on the side portfolio, which in turn allows for a higher VPW withdrawal rate for the side portfolio.

So, in essence my question is how does a traditional SWR withdrawal compare to A TIPS floor and an 80:20 VPW. I am going to check into this in more detail. I think this can be quantified.


The classic financial solution is to annuitize everything. People seem to not like that very much.

Annuity is typically not an option due to liquidity. My plan only involves an SPIA at age 80.

As far as sequence of returns, there are two uncertainties. One is the sequence of a known average of returns. The other is uncertainty as to the unknown future average. RTM involves the question of return to what mean. Aside from that, the data one assumes has to be real returns and/or one has to also have a sequence of inflation in mind. If sources of income and expenses mix real and nominal instruments, estimating sequence of inflation cannot be avoided.

It's only the unknown sequence of a known average real returns that's unique to retirement though.

On top of this actual retirements don't have fixed needs for income. A retiree has to estimate the variability and sequence of expenses. I think overengineering retirement is a siren song.

Income flexibility is ignored. Everyone has an optimal retirement life in mind, sure you could easily lower it, but IMO that's not success. Success is enjoying an income you had pre-determined...If I have to lower it, by definition I am not enjoying my planned retirement. I'm not going to cry over it. But, I am planning to 'assure' that income. Therefore flexibility is ignored.

gilgamesh
Posts: 835
Joined: Sun Jan 10, 2016 9:29 am

Re: Safemax with varying AA

Post by gilgamesh » Thu Jun 15, 2017 5:19 am

AlohaJoe wrote:
Or did you do all 30 year periods from 1926-2010 or some similar range?
Yes, this is what I did. I used Shiller's annual data which covers 1871-2016.
Where did you get the inflation data? Is that an annual average? Thanks!

P.S: Can you expand on exactly how you did this?

AlohaJoe
Posts: 2423
Joined: Mon Nov 26, 2007 2:00 pm
Location: Saigon, Vietnam

Re: Safemax with varying AA

Post by AlohaJoe » Thu Jun 15, 2017 5:29 am

gilgamesh wrote:
AlohaJoe wrote:
Or did you do all 30 year periods from 1926-2010 or some similar range?
Yes, this is what I did. I used Shiller's annual data which covers 1871-2016.
Where did you get the inflation data? Is that an annual average? Thanks!
Shiller's data includes inflation, no it isn't an annual average. You can read it here: http://www.econ.yale.edu/~shiller/data.htm
P.S: Can you expand on exactly how you did this?
For a given portfolio, you calculate the annual returns for that year. For a portfolio that is 50/50 that just means you take the average of equity & bond returns for that year. Once you have a sequence of portfolio returns, there's a formula for calculating the maximum you can withdraw

Image

I've written some software that does it all for me.

https://github.com/hoostus/prime-harves ... 0%25.ipynb

gilgamesh
Posts: 835
Joined: Sun Jan 10, 2016 9:29 am

Re: Safemax with varying AA

Post by gilgamesh » Thu Jun 15, 2017 7:25 am

AlohaJoe wrote:
gilgamesh wrote:
AlohaJoe wrote:
Or did you do all 30 year periods from 1926-2010 or some similar range?
Yes, this is what I did. I used Shiller's annual data which covers 1871-2016.
Where did you get the inflation data? Is that an annual average? Thanks!
Shiller's data includes inflation, no it isn't an annual average. You can read it here: http://www.econ.yale.edu/~shiller/data.htm
P.S: Can you expand on exactly how you did this?
For a given portfolio, you calculate the annual returns for that year. For a portfolio that is 50/50 that just means you take the average of equity & bond returns for that year. Once you have a sequence of portfolio returns, there's a formula for calculating the maximum you can withdraw

Image

I've written some software that does it all for me.

https://github.com/hoostus/prime-harves ... 0%25.ipynb
It's been over 20 years since I've dealt with analysis of this sort, my career had taken a different path. It sounds like you definitely know what you are doing...thanks!

Can you answer this. Today, can we assume the difference between the yield on a nominal treasury bond and TIPS of the same duration is the expected inflation? I might be getting some terms wrong here, if so, what among these is the Feds expected inflation into the future when comparing equivalent nominal trasuries and TIPS?

Now, do the Feds release any other indicators of their expected future inflation? If they had done so in the past, and knowing the nominal bond rates of the past, we could mimic TIPS well into the past when they were never offered. Is there such a measure?

Thanks!
Last edited by gilgamesh on Thu Jun 15, 2017 8:17 am, edited 2 times in total.

gilgamesh
Posts: 835
Joined: Sun Jan 10, 2016 9:29 am

Re: Safemax with varying AA

Post by gilgamesh » Thu Jun 15, 2017 7:47 am

MIpreRetirey wrote:
gilgamesh wrote:
MIpreRetirey wrote:
gilgamesh wrote:
itstoomuch wrote:Finally, once one achieves a secure FR, one can do any thing with the excess assets. For me, I acquired more equities in Discretionay acct (rising equity portfolio in retirement, see threads on this). :greedy
I of course have a similar plan, and I'm trying to see how it fits with the traditional SAFEMAX SWR concept.

Will flooring with TIPS from retirement age of 60 to SS at age 70 (happens to be the first 10 years of retirement) also result in higher SWR, as viewed through the traditional SWR withdrawal.
I'd look at the flooring is still part of your total portfolio as far as total WR. But, you can consider it a source of inflation adjusted income (after subtracting it's initial total cost from your total portfolio) in firecalc.com. On the last tab in firecalc, you can investigate the best AA for your WR and time needed, and other things.
I often run two firecalcs: one for pre-SS, and then one for a new WR after SS.
You know what, under the 'Other income/spending' tab, I could use the CPI adjusted pension for the TIPS floor. It will however ignore the yield which I can approximate elsewhere too. Didn't realize firecalc is this robust.

But, I still need to figure out how to mimic TIPS that vary over years to that of a pension. TIPS will be certain amount for the first ten years, then lower for the next 20 with SS and then will be zero to be replaced by an SPIA. If anyone has any tips, that'll be great, but otherwise I should be able to figure it out by trial and error. The instructions for pension is a little murky to me.
I think you won't be able to model TIPS very well for more than 20 years backlooking. See these two threads for efforts on TIPS: (search "TIPS" ?)
viewtopic.php?t=120430
viewtopic.php?t=2520
The links talk about VPW. IMO, this is why it's worse than SWR, for what I am trying to achieve. My aim is to spend as much as possible for the first ten years of retirement from ages 60-70 when I will travel most, but still 'assuring' future income.

VPW reacts instantly to market condition. Say if my portfolio goes down 30% on my second year of retirement then my withdrawals will go down as well. If it goes up on my 15th year, then my income goes up for that year. The WR is at the mercy of the markets, not when I want to spend more.

At least, SWR ignores current market conditions, and blindly follows a predetermined WR, just with the knowledge of the past and knowing it's always worked out. So, even if there were a market crash on the 3rd year my income will not go down.

Also, IMO, VPW will be very sensitive to prolonged high inflation. Unlike SWR, it ignores inflation, and that is a big problem too.

Again. I could be wrong, and if I am, please let me know. As the first ten years are sensitive to sequence risk, and SWR has to survive a full 30 years, SWR has to be untra conservative to traverse the unknowns of the first 10 years. Then, most of the times nothing undue happens in the first ten years, and a retiree is left with a huge nest egg precisely when they don't need much.

Can this spending me evened out, where a retiree can spend more, and more comfortably knowing future income is safe, during the first ten years of retirement?

Annuity is not an option. Assume no other income like pension etc.

How about a TIPS ladder for the first 10 years covering 'X' amount of retirement income needed ? We are hedging inflation and sequence risks. A side portfolio with 80/20 with a VPW withdrawal convering the (income needed - X).

After age 70, SS will take up most of the floor income, however if X>SS, the difference is made up by another TIPS ladder equaling X-SS.

This IMO should allow for worry free spending for the first 10 years, irrelevant of market condition. As my future income is mostly guaranteed.

Then I want to figure out, whether I am losing anything for this luxury? I am sure I will lose out on the potential upside by going more bond heavy. But, to me, I want to know whether such a set up will allow me to spend more in the first 10 years of retirement than traditional SWR based income strategy?

Now, even if someone has no desire to set up a floor for X amount. It still applies....everyone has a certain bond allocation, just use all that bond allocation constructing a TIPS ladder. This will apply universally. This to me is hyper interesting. SWR with the traditional stock/bond of 60:40. Or all 40% in a TIPS ladder (most of which will be used from retirement to SS) and remaining 60% invested at 80/20 and follows VPW. How do these two compare?...This would require extrapolating TIPS data into the past when no TIPS existed, thus my previous post in trying to figure out a solution to that part of the problem.

It's easy to run this for my particular situation, as I know what my projected numbers are, and I will do that over the weekend.

In the mean time, I am posting this to see whether what I am trying to say makes sense? Or is it all nonsense, with no substance?

P.S: I would love to hear from both sides. Please let me know, if it kind of make sense to you...also, please let me know if all of this is non-sensical.

AlohaJoe
Posts: 2423
Joined: Mon Nov 26, 2007 2:00 pm
Location: Saigon, Vietnam

Re: Safemax with varying AA

Post by AlohaJoe » Thu Jun 15, 2017 9:21 am

gilgamesh wrote:Can you answer this. Today, can we assume the difference between the yield on a nominal treasury bond and TIPS of the same duration is the expected inflation?
No. The difference between nominal Treasuries and TIPS is the "break-even inflation". But that's not the same as "expected inflation". Break-even inflation = expected inflation + risk premium. Since the risk premium varies over time, you can't use the break-even inflation as a measure of expected future inflation.
we could mimic TIPS well into the past when they were never offered. Is there such a measure?
A link provided up-thread chronicles the (failed) efforts to build a historical TIPS dataset. No, there is currently no way to mimic TIPS well into the past.

gilgamesh
Posts: 835
Joined: Sun Jan 10, 2016 9:29 am

Re: Safemax with varying AA

Post by gilgamesh » Thu Jun 15, 2017 9:59 am

AlohaJoe wrote:
gilgamesh wrote:Can you answer this. Today, can we assume the difference between the yield on a nominal treasury bond and TIPS of the same duration is the expected inflation?
No. The difference between nominal Treasuries and TIPS is the "break-even inflation". But that's not the same as "expected inflation". Break-even inflation = expected inflation + risk premium. Since the risk premium varies over time, you can't use the break-even inflation as a measure of expected future inflation.

Yup! Makes sense....Thanks!
we could mimic TIPS well into the past when they were never offered. Is there such a measure?
A link provided up-thread chronicles the (failed) efforts to build a historical TIPS dataset. No, there is currently no way to mimic TIPS well into the past.

I think this is the thread, but wasn't mentioned above though

viewtopic.php?f=10&t=203092&p=3160853#p3159772

.

MIpreRetirey
Posts: 632
Joined: Fri Sep 06, 2013 12:35 pm

Re: Safemax with varying AA

Post by MIpreRetirey » Thu Jun 15, 2017 9:39 pm

gilgamesh wrote:
MIpreRetirey wrote:
gilgamesh wrote:
MIpreRetirey wrote:
gilgamesh wrote:
I of course have a similar plan, and I'm trying to see how it fits with the traditional SAFEMAX SWR concept.

Will flooring with TIPS from retirement age of 60 to SS at age 70 (happens to be the first 10 years of retirement) also result in higher SWR, as viewed through the traditional SWR withdrawal.
I'd look at the flooring is still part of your total portfolio as far as total WR. But, you can consider it a source of inflation adjusted income (after subtracting it's initial total cost from your total portfolio) in firecalc.com. On the last tab in firecalc, you can investigate the best AA for your WR and time needed, and other things.
I often run two firecalcs: one for pre-SS, and then one for a new WR after SS.
You know what, under the 'Other income/spending' tab, I could use the CPI adjusted pension for the TIPS floor. It will however ignore the yield which I can approximate elsewhere too. Didn't realize firecalc is this robust.

But, I still need to figure out how to mimic TIPS that vary over years to that of a pension. TIPS will be certain amount for the first ten years, then lower for the next 20 with SS and then will be zero to be replaced by an SPIA. If anyone has any tips, that'll be great, but otherwise I should be able to figure it out by trial and error. The instructions for pension is a little murky to me.
I think you won't be able to model TIPS very well for more than 20 years backlooking. See these two threads for efforts on TIPS: (search "TIPS" ?)
viewtopic.php?t=120430
viewtopic.php?t=2520
The links talk about VPW. IMO, this is why it's worse than SWR, for what I am trying to achieve. My aim is to spend as much as possible for the first ten years of retirement from ages 60-70 when I will travel most, but still 'assuring' future income.

VPW reacts instantly to market condition. Say if my portfolio goes down 30% on my second year of retirement then my withdrawals will go down as well. If it goes up on my 15th year, then my income goes up for that year. The WR is at the mercy of the markets, not when I want to spend more.

At least, SWR ignores current market conditions, and blindly follows a predetermined WR, just with the knowledge of the past and knowing it's always worked out. So, even if there were a market crash on the 3rd year my income will not go down.

Also, IMO, VPW will be very sensitive to prolonged high inflation. Unlike SWR, it ignores inflation, and that is a big problem too.

Again. I could be wrong, and if I am, please let me know. As the first ten years are sensitive to sequence risk, and SWR has to survive a full 30 years, SWR has to be untra conservative to traverse the unknowns of the first 10 years. Then, most of the times nothing undue happens in the first ten years, and a retiree is left with a huge nest egg precisely when they don't need much.

Can this spending me evened out, where a retiree can spend more, and more comfortably knowing future income is safe, during the first ten years of retirement?

Annuity is not an option. Assume no other income like pension etc.

How about a TIPS ladder for the first 10 years covering 'X' amount of retirement income needed ? We are hedging inflation and sequence risks. A side portfolio with 80/20 with a VPW withdrawal convering the (income needed - X).

After age 70, SS will take up most of the floor income, however if X>SS, the difference is made up by another TIPS ladder equaling X-SS.

This IMO should allow for worry free spending for the first 10 years, irrelevant of market condition. As my future income is mostly guaranteed.

Then I want to figure out, whether I am losing anything for this luxury? I am sure I will lose out on the potential upside by going more bond heavy. But, to me, I want to know whether such a set up will allow me to spend more in the first 10 years of retirement than traditional SWR based income strategy?

Now, even if someone has no desire to set up a floor for X amount. It still applies....everyone has a certain bond allocation, just use all that bond allocation constructing a TIPS ladder. This will apply universally. This to me is hyper interesting. SWR with the traditional stock/bond of 60:40. Or all 40% in a TIPS ladder (most of which will be used from retirement to SS) and remaining 60% invested at 80/20 and follows VPW. How do these two compare?...This would require extrapolating TIPS data into the past when no TIPS existed, thus my previous post in trying to figure out a solution to that part of the problem.

It's easy to run this for my particular situation, as I know what my projected numbers are, and I will do that over the weekend.

In the mean time, I am posting this to see whether what I am trying to say makes sense? Or is it all nonsense, with no substance?

P.S: I would love to hear from both sides. Please let me know, if it kind of make sense to you...also, please let me know if all of this is non-sensical.
Not non-sensical in my view at all. It's been my starting approach, and still generally my way, what with the step in income upon taking SS.
My two links, yes were the VPW and Simba backtest data threads, but I meant to try a search for 'TIPS' at the top box on those threads. Much effort and discussion about the 'no historic data on TIPS'. I think most discussion was c. 2015, 2016?

My opinion on the VPW method is:
it's root method is a WR model of 1/n, n years left in retirement plan. ie: seq. of WD: 1/40th, 1/39th..1/1 last year (ALL spent that last year if you live that long.)
It's starts with assumptions or expectations which you can change in the spread sheet about future returns of stocks and bonds.
It builds an initial table of the factor you use for WD each year. (similar to an RMD table that starts at age 70.5?)

But if you change the expected return in the formula to something higher, it will 'front load' your WDs to the beginning years, until the ACTUAL returns, if less, you are getting squishes down your portfolio somewhat more than the % return you used as an input.

At least that is my understanding. But it would tend to be a smoother transition, all things equal, than planning with abrupt changes.

So, even with VPW, and most approaches, I tend to side with dividing my retirement planning into 2 time periods; one pre-SS, 1 for after.
You can make a simple equation that allows equal spending with the step of SS income included.
I think looking at retirement spending each year is better to look from the end of retirement and then backwards to the start.
It makes the formula easier to create. Some threads have mentioned this, too, with formulas, last couple years or so. (#cruncher posts)
I can't answer much with detail.
new username.

itstoomuch
Posts: 4554
Joined: Mon Dec 15, 2014 12:17 pm
Location: midValley OR

Re: Safemax with varying AA

Post by itstoomuch » Thu Jun 15, 2017 10:24 pm

aloha Joe wrote: {...snip...}
For a given portfolio, you calculate the annual returns for that year. For a portfolio that is 50/50 that just means you take the average of equity & bond returns for that year. Once you have a sequence of portfolio returns, there's a formula for calculating the maximum you can withdraw

Image

I've written some software that does it all for me.

https://github.com/hoostus/prime-harves ... 0%25.ipynb
Looks like a distributed network of capacitors in series.
Rev90517; 4 Incm stream buckets: SS+pension; dfr'd GLWB VA & FI anntys, by time & $$ laddered; Discretionary; Rentals. LTCi. Own, not asset. Tax 25%. Early SS. FundRatio (FR) >1.1 67/70yo

gilgamesh
Posts: 835
Joined: Sun Jan 10, 2016 9:29 am

Re: Safemax with varying AA

Post by gilgamesh » Fri Jun 16, 2017 6:56 am

MIpreRetirey wrote: Not non-sensical in my view at all. It's been my starting approach, and still generally my way, what with the step in income upon taking SS.
My two links, yes were the VPW and Simba backtest data threads, but I meant to try a search for 'TIPS' at the top box on those threads. Much effort and discussion about the 'no historic data on TIPS'. I think most discussion was c. 2015, 2016?

My opinion on the VPW method is:
it's root method is a WR model of 1/n, n years left in retirement plan. ie: seq. of WD: 1/40th, 1/39th..1/1 last year (ALL spent that last year if you live that long.)
It's starts with assumptions or expectations which you can change in the spread sheet about future returns of stocks and bonds.
It builds an initial table of the factor you use for WD each year. (similar to an RMD table that starts at age 70.5?)

But if you change the expected return in the formula to something higher, it will 'front load' your WDs to the beginning years, until the ACTUAL returns, if less, you are getting squishes down your portfolio somewhat more than the % return you used as an input.

At least that is my understanding. But it would tend to be a smoother transition, all things equal, than planning with abrupt changes.

So, even with VPW, and most approaches, I tend to side with dividing my retirement planning into 2 time periods; one pre-SS, 1 for after.
You can make a simple equation that allows equal spending with the step of SS income included.
I think looking at retirement spending each year is better to look from the end of retirement and then backwards to the start.
It makes the formula easier to create. Some threads have mentioned this, too, with formulas, last couple years or so. (#cruncher posts)
I can't answer much with detail.
If the spreadsheet allows to input your own returns for stocks and bonds then I assume the expected return in the spreadsheet is to figure out the actual percentage to withdraw. Say, if you put an expected return of 100% each year for stocks then the percentage you can withdraw will be higher. I guess one could input an expected return higher to skew the results to withdraw more initially...eventually the actual rate will catch up though.

I did not use the spreadsheet, but the tables below should be identical to the spreadsheet for the ages and AA. There is no way to input expected return on the table. They may have used the past performance though. We are just withdrawing a certain percentage from the existing balance. So, here only the actual returns matter. If the actual returns matches the past, the income will be distributed evenly, on years where it's higher we'll have higher income, when it's lower we'll have less retirement income. But the percentage each year is calculated assuming past averages for the remainder of the year.

However, anything other than the historic past IMO is not prudent...if anything one may need to go more conservative. In the accumulation phase, I can input the average returns and be ok, in retirement average doesn't matter due to sequence risk. So, artificially increasing the returns higher than the past is not what I want to do.

Why do you think the tables doesn't have the option to change returns...given that, do you think my interpretation above is correct?

Thank you!

MIpreRetirey
Posts: 632
Joined: Fri Sep 06, 2013 12:35 pm

Re: Safemax with varying AA

Post by MIpreRetirey » Fri Jun 16, 2017 7:13 am

gilgamesh wrote:
MIpreRetirey wrote: Not non-sensical in my view at all. It's been my starting approach, and still generally my way, what with the step in income upon taking SS.
My two links, yes were the VPW and Simba backtest data threads, but I meant to try a search for 'TIPS' at the top box on those threads. Much effort and discussion about the 'no historic data on TIPS'. I think most discussion was c. 2015, 2016?

My opinion on the VPW method is:
it's root method is a WR model of 1/n, n years left in retirement plan. ie: seq. of WD: 1/40th, 1/39th..1/1 last year (ALL spent that last year if you live that long.)
It's starts with assumptions or expectations which you can change in the spread sheet about future returns of stocks and bonds.
It builds an initial table of the factor you use for WD each year. (similar to an RMD table that starts at age 70.5?)

But if you change the expected return in the formula to something higher, it will 'front load' your WDs to the beginning years, until the ACTUAL returns, if less, you are getting squishes down your portfolio somewhat more than the % return you used as an input.

At least that is my understanding. But it would tend to be a smoother transition, all things equal, than planning with abrupt changes.

So, even with VPW, and most approaches, I tend to side with dividing my retirement planning into 2 time periods; one pre-SS, 1 for after.
You can make a simple equation that allows equal spending with the step of SS income included.
I think looking at retirement spending each year is better to look from the end of retirement and then backwards to the start.
It makes the formula easier to create. Some threads have mentioned this, too, with formulas, last couple years or so. (#cruncher posts)
I can't answer much with detail.
If the spreadsheet allows to input your own returns for stocks and bonds then I assume the expected return in the spreadsheet is to figure out the actual percentage to withdraw. Say, if you put an expected return of 100% each year for stocks then the percentage you can withdraw will be higher. I guess one could input an expected return higher to skew the results to withdraw more initially...eventually the actual rate will catch up though.

I did not use the spreadsheet, but the tables below should be identical to the spreadsheet for the ages and AA. There is no way to input expected return on the table. They may have used the past performance though. We are just withdrawing a certain percentage from the existing balance. So, here only the actual returns matter. If the actual returns matches the past, the income will be distributed evenly, on years where it's higher we'll have higher income, when it's lower we'll have less retirement income. But the percentage each year is calculated assuming past averages for the remainder of the year.

However, anything other than the historic past IMO is not prudent...if anything one may need to go more conservative. In the accumulation phase, I can input the average returns and be ok, in retirement average doesn't matter due to sequence risk. So, artificially increasing the returns higher than the past is not what I want to do.

Why do you think the tables doesn't have the option to change returns...given that, do you think my interpretation above is correct?

Thank you!
Of course you are correct. But think about it. It sets up a table of % of portfolio to WD every year. If you don't get the expected returns, then the port. drops in value faster, even though the table of %WDs just takes a % of what's left. A steeper ski slope than, would of been an attempt at just a deadline for complete consumption at an effort at constant dollar WD .

I thought from use b4 that if you download the spreadsheet you can modify unprotected variables.??
new username.

User avatar
Nearing_Destination
Posts: 677
Joined: Wed Aug 05, 2009 1:26 pm

Re: Safemax with varying AA

Post by Nearing_Destination » Sun Jun 18, 2017 3:53 pm

itstoomuch wrote:
aloha Joe wrote: {...snip...}
For a given portfolio, you calculate the annual returns for that year. For a portfolio that is 50/50 that just means you take the average of equity & bond returns for that year. Once you have a sequence of portfolio returns, there's a formula for calculating the maximum you can withdraw

Image

I've written some software that does it all for me.

https://github.com/hoostus/prime-harves ... 0%25.ipynb
Looks like a distributed network of capacitors in series.
Should act like one too....

Wouldn't the problem be supplying the overvoltage ($$$) needed to charge it?

MIpreRetirey
Posts: 632
Joined: Fri Sep 06, 2013 12:35 pm

Re: Safemax with varying AA

Post by MIpreRetirey » Sun Jun 18, 2017 4:29 pm

Nearing_Destination wrote:
itstoomuch wrote:
aloha Joe wrote: {...snip...}
For a given portfolio, you calculate the annual returns for that year. For a portfolio that is 50/50 that just means you take the average of equity & bond returns for that year. Once you have a sequence of portfolio returns, there's a formula for calculating the maximum you can withdraw

Image

I've written some software that does it all for me.

https://github.com/hoostus/prime-harves ... 0%25.ipynb
Looks like a distributed network of capacitors in series.
Should act like one too....

Wouldn't the problem be supplying the overvoltage ($$$) needed to charge it?
"More importantly, we have a charge", 'Real Genius'
new username.

Post Reply