Total Portfolio Allocation and Withdrawal (TPAW)

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
Circle the Wagons
Posts: 164
Joined: Mon Jan 15, 2024 1:34 pm

Re: Total Portfolio Allocation and Withdrawal (TPAW)

Post by Circle the Wagons »

GAAP wrote: Mon Jul 08, 2024 1:25 pm
ConstantChrysalis wrote: Sun Jul 07, 2024 9:37 pm
Raspberry-503 wrote: Sun Jul 07, 2024 9:23 pm This is also where picking options for future stocks valuations makes a huge different, 1/CAPE and historical will make a huge difference, so there is still a lot of guesswork behind the formulas.
This is a good example. Picking historical will give us a more familiar and aggressive equity asset allocation. But, I've read works by several smart people who persuaded me that we will likely experience lower equity returns than those historical results going forward. 1/CAPE yields a less aggressive equity asset allocation and there is a fair amount of debate about how good it is at prediction. After weighing these two, I chose to go with the 1/CAPE variation that TPAW uses as its default, and accepted the consequence of the lower equity allocation. Someone else might come to a different conclusion and pick historical. My caution here is that we shouldn't just blindly pick one or the other just because the allocation turns out like we want.
1/CAPE also adds the notable advantage of changing as conditions change rather than hoping that conditions will be better than the historical norm (and not worse). For that reason, it is particularly appropriate for a plan that is revisited regularly -- as all good plans should be.
Can folks share if they're using pure 1/CAPE or one of the regression variations, and perhaps why? The former is very conservative, so this decision is big for the plan.

Personally, I use the Conservative Estimate based on regressions. For the avoidance of doubt, this has an ERP of 1.6% as of yesterday. It feels appropriately conservative while still being within arm's length of other long-term projections, such as Vanguard's.

For pure 1/CAPE, the ERP is only 0.6%.
GAAP
Posts: 2784
Joined: Fri Apr 08, 2016 12:41 pm

Re: Total Portfolio Allocation and Withdrawal (TPAW)

Post by GAAP »

1/CAPE is a power-curve approximation of a growth function that is generally logarithmic. The results estimated by 1/CAPE are reasonably similar to a logarithmic regression function for CAPE values in the range of about 8 or 9 to 30-35. Outside that range 1/CAPE is a pretty poor match -- too optimistic at both extremes, and with no possibility of negative growth no matter how high the CAPE value. It's the lower extreme that matters the most, since 1/CAPE will generate hopelessly optimistic estimates. At the upper extreme, it's the difference between tiny and tinier returns.

Perfect correlation isn't really necessary, because the plan will adjust to performance over time. You just need to be in the ballpark.
“Adapt what is useful, reject what is useless, and add what is specifically your own.” ― Bruce Lee
ConstantChrysalis
Posts: 82
Joined: Sun Aug 15, 2021 5:53 pm

Re: Total Portfolio Allocation and Withdrawal (TPAW)

Post by ConstantChrysalis »

I was too loose when saying I used 1/CAPE. Specifically, I'm using the default which is Regression Prediction.
Circle the Wagons
Posts: 164
Joined: Mon Jan 15, 2024 1:34 pm

Re: Total Portfolio Allocation and Withdrawal (TPAW)

Post by Circle the Wagons »

GAAP wrote: Mon Jul 08, 2024 2:56 pm 1/CAPE is a power-curve approximation of a growth function that is generally logarithmic. The results estimated by 1/CAPE are reasonably similar to a logarithmic regression function for CAPE values in the range of about 8 or 9 to 30-35. Outside that range 1/CAPE is a pretty poor match -- too optimistic at both extremes, and with no possibility of negative growth no matter how high the CAPE value. It's the lower extreme that matters the most, since 1/CAPE will generate hopelessly optimistic estimates. At the upper extreme, it's the difference between tiny and tinier returns.
Thank you. Am I correct in interpreting the implication of your post as follows: 1/CAPE is too optimistic at CAPE extremes. Since CAPE is over 36 as we speak, it is too optimistic today. Therefore, 1/CAPE, even as the most conservative option available for the ERP assumption in TPAW today, is still too optimistic ... ? ... are you hardwiring in something even more conservative, like an ER for stocks that is below the 20-yr TIPS yield?
Perfect correlation isn't really necessary, because the plan will adjust to performance over time. You just need to be in the ballpark.
Appreciate this but in setting a baseline AA, there is a huge difference in what TPAW recommends depending on the ERP assumptions chosen.
User avatar
Raspberry-503
Posts: 1106
Joined: Sat Oct 03, 2020 6:42 am

Re: Total Portfolio Allocation and Withdrawal (TPAW)

Post by Raspberry-503 »

Circle the Wagons wrote: Mon Jul 08, 2024 2:14 pm Can folks share if they're using pure 1/CAPE or one of the regression variations, and perhaps why? The former is very conservative, so this decision is big for the plan.
CAPE ratio today is 36.29 (here it seems that I can find multiple numbers depending on source)
1/36.29 is 2.76%
TPAW today show stock returns for "1/CAPE" as 3.7% EDIT: OOPS misread it, it is indeed 2.7%)


EDIT: 1/CAPE is also for US markets (or more exactly the CAPE will be different for international markets)

EDIT: longinvest, one of the authors of the VPW withdrawal method, insists it's a fallacy to try and predict returns, and you might as well pick a wild-ass-guess that never changes. That WAGis 5% stock and 1.9% bonds for VPW, so not far from the defaults used by TPAW.
Last edited by Raspberry-503 on Mon Jul 08, 2024 6:26 pm, edited 1 time in total.
Circle the Wagons
Posts: 164
Joined: Mon Jan 15, 2024 1:34 pm

Re: Total Portfolio Allocation and Withdrawal (TPAW)

Post by Circle the Wagons »

Raspberry-503 wrote: Mon Jul 08, 2024 5:36 pm
Circle the Wagons wrote: Mon Jul 08, 2024 2:14 pm Can folks share if they're using pure 1/CAPE or one of the regression variations, and perhaps why? The former is very conservative, so this decision is big for the plan.
CAPE ratio today is 36.29 (here it seems that I can find multiple numbers depending on source)
1/36.29 is 2.76%
TPAW today show stock returns for "1/CAPE" as 3.7%

so it's not using "1 divided by the current CAPE ratio".

It seems in line with this article and the table at the bottom: https://www.advisorperspectives.com/art ... -returns-1 but they don't spell out the curve-fitting parameters.


EDIT: 1/CAPE is also for US markets (or more exactly the CAPE will be different for international markets)

EDIT: longinvest, one of the authors of the VPW withdrawal method, insists it's a fallacy to try and predict returns, and you might as well pick a wild-ass-guess that never changes. That WAGis 5% stock and 1.9% bonds for VPW, so not far from the defaults used by TPAW.
The default regression prediction is 5.1%.
The conservative regression is 3.7%.
The pure 1/CAPE in TPAW is indeed 2.7%. Minus 2.1% for the TIPS 20-yr gives you ERP of 0.6%.
User avatar
Raspberry-503
Posts: 1106
Joined: Sat Oct 03, 2020 6:42 am

Re: Total Portfolio Allocation and Withdrawal (TPAW)

Post by Raspberry-503 »

Circle the Wagons wrote: Mon Jul 08, 2024 6:12 pm The pure 1/CAPE in TPAW is indeed 2.7%.
Yes, my bad, I read the wrong value earlier, so yes, it really is "1 divided by CAPE", I edited my post above.

I also missed that TAPW has a pop up explanation about those calculations in the Expected Returns selection section, so it tells you right there.

International CAPE is much lower, but I don't known that 1/CAPE applies there (I don't see why it should be different?), I'm not sure where to find numbers but I remember seeing 10-12 for Emerging and 20-25 for Developed.
Circle the Wagons
Posts: 164
Joined: Mon Jan 15, 2024 1:34 pm

Re: Total Portfolio Allocation and Withdrawal (TPAW)

Post by Circle the Wagons »

Raspberry-503 wrote: Mon Jul 08, 2024 6:32 pm
Circle the Wagons wrote: Mon Jul 08, 2024 6:12 pm The pure 1/CAPE in TPAW is indeed 2.7%.
Yes, my bad, I read the wrong value earlier, so yes, it really is "1 divided by CAPE", I edited my post above.

I also missed that TAPW has a pop up explanation about those calculations in the Expected Returns selection section, so it tells you right there.

International CAPE is much lower, but I don't known that 1/CAPE applies there (I don't see why it should be different?), I'm not sure where to find numbers but I remember seeing 10-12 for Emerging and 20-25 for Developed.
Right. Int'l is one of the reasons I got more comfortable with the Conservative estimate vs. pure 1/CAPE. I believe Ben has also cited it as one of the reasons he moved the default up from the Conservative regression a few months back.
ConstantChrysalis
Posts: 82
Joined: Sun Aug 15, 2021 5:53 pm

Re: Total Portfolio Allocation and Withdrawal (TPAW)

Post by ConstantChrysalis »

I often wonder if the specific metric here just doesn't matter much in the big picture, and the real value is in the disciplined and mechanical rebalancing that helps one avoid behavioral errors and generally encourages buying low and selling high. Maybe, similar to having a very wide range of static asset allocations that work approximately equally well for most people (like 30/70 to 70/30) with disciplined rebalancing yet worrying about dialing it into the nearest 1%.
User avatar
Raspberry-503
Posts: 1106
Joined: Sat Oct 03, 2020 6:42 am

Re: Total Portfolio Allocation and Withdrawal (TPAW)

Post by Raspberry-503 »

ConstantChrysalis wrote: Mon Jul 08, 2024 6:58 pm I often wonder if the specific metric here just doesn't matter much in the big picture, and the real value is in the disciplined and mechanical rebalancing that helps one avoid behavioral errors and generally encourages buying low and selling high. Maybe, similar to having a very wide range of static asset allocations that work approximately equally well for most people (like 30/70 to 70/30) with disciplined rebalancing yet worrying about dialing it into the nearest 1%.
I'm thinking along the same lines (and the VPW embraces that). I think a way to approximate that is to set SPAW and set your AA to different values and see how much it changes the montecarlo simulation.
Circle the Wagons
Posts: 164
Joined: Mon Jan 15, 2024 1:34 pm

Re: Total Portfolio Allocation and Withdrawal (TPAW)

Post by Circle the Wagons »

Raspberry-503 wrote: Mon Jul 08, 2024 7:08 pm
ConstantChrysalis wrote: Mon Jul 08, 2024 6:58 pm I often wonder if the specific metric here just doesn't matter much in the big picture, and the real value is in the disciplined and mechanical rebalancing that helps one avoid behavioral errors and generally encourages buying low and selling high. Maybe, similar to having a very wide range of static asset allocations that work approximately equally well for most people (like 30/70 to 70/30) with disciplined rebalancing yet worrying about dialing it into the nearest 1%.
I'm thinking along the same lines (and the VPW embraces that). I think a way to approximate that is to set SPAW and set your AA to different values and see how much it changes the montecarlo simulation.
To play devil's advocate here, one of the (many) reasons I really like TPAW is it can help the investor objectively process implications of a fluctuating ERP, even and especially in extreme situations that might push outside of an assumed "reasonable range" like 30-70% equity. But the ER assumptions chosen matter a lot -- in my case, anywhere from 27-100% suggested stocks just from the first three choices. This is why I raised the topic and appreciate hearing how folks think about it.
User avatar
Raspberry-503
Posts: 1106
Joined: Sat Oct 03, 2020 6:42 am

Re: Total Portfolio Allocation and Withdrawal (TPAW)

Post by Raspberry-503 »

I switched to SPAW and compared 30/70 and 7/30 (and default returns), the graph doesn't show it but expect for the first 2 years, all essential expenses (what I call the dignity floor, but not enough for travel etc...) are covered by SS or a TIPS ladder. so the portfolio only funds non-essential expenses, and I don't care to leave a legacy. That means that I'm not super concerned about variations in the portfolio, especially year-to-year (during a market crash)

Image

TPAW recommends 39/61 and falls in the middle
Image

Picking 1/CAPE for grim returns (although bonds are pretty good), definitely yields lower upside, with 8/92 AA
Image

So in this case, while the AA varies widely, the projected lower income varies around 5% between best and worse case, greater returns will provide grater upside, so will a more aggressive AA.

I'm still processing what it might mean, but in my case I think I don't think "getting it wrong" is catastrophic.
User avatar
Topic Author
Ben Mathew
Posts: 2819
Joined: Tue Mar 13, 2018 11:41 am
Location: Seattle

Re: Total Portfolio Allocation and Withdrawal (TPAW)

Post by Ben Mathew »

jmk wrote: Sun Jul 07, 2024 6:08 pm I had erroneously assumed that both the future Expenses and future Income were being discounted at the same rate (20y TIP). So that while the AAs through time would change dynamically, the initial AA would not be affected by the equity ER at all. I had assumed NPV (Tip rate, years, Expense steam) = NPV (Tip Rate, years, Income stream +Portfolio at RRA). But I can see you are correct: fiddling with the equity ER assumptions radically affects the AA, even on the very first withdrawal. Is this solely because of the Monte Carlo part of TPAW for the Discretionary part of the portfolio ("General Spending") which is of course affected by equity ER?
Extra essential expenses are funded with 100% bonds and so discounted by the TIPS rate. Extra discretionary expenses and general expenses are funded with the risk portfolio. The asset allocation of the risk portfolio is determined by Merton's formula and so will depend on the equity risk premium.

So unless extra essential expenses use up all of wealth (making the asset allocation 100% bonds), there will be a risk portfolio and the asset allocation will depend on the equity risk premium.
Total Portfolio Allocation and Withdrawal (TPAW)
User avatar
Topic Author
Ben Mathew
Posts: 2819
Joined: Tue Mar 13, 2018 11:41 am
Location: Seattle

Re: Total Portfolio Allocation and Withdrawal (TPAW)

Post by Ben Mathew »

ConstantChrysalis wrote: Sun Jul 07, 2024 8:58 pm Most planners let you pick an asset allocation as in input assumption. The TPAW approach is to output a recommended asset allocation based on current equity and bond valuations and volatilities. [...]

I think there is a trap behind this philosophy difference between most tools and TPAW. It's easy to change the assumptions until they match what you expect, like 60/40 or 70/30. [...]
Yes, changing assumptions to match a target asset allocation defeats the purpose of the planner. The goal is to discover the asset allocation that's right for you, which depends on your risk preferences as well as the equity risk premium and equity risk. The way to do that is to stay focused on the "Monthly Spending During Retirement" graph when selecting the asset allocation because that's what communicates the risk vs reward offered by stocks vs bonds.

Asset allocation should be an output of the planning process, not an input!
Total Portfolio Allocation and Withdrawal (TPAW)
User avatar
Topic Author
Ben Mathew
Posts: 2819
Joined: Tue Mar 13, 2018 11:41 am
Location: Seattle

Re: Total Portfolio Allocation and Withdrawal (TPAW)

Post by Ben Mathew »

ConstantChrysalis wrote: Mon Jul 08, 2024 6:46 am After spending some time on this, I’ve come away with an even greater appreciation for what the Matthew brothers have done for us.
Thanks for the kind words. It's exciting to be able to contribute towards helping popularize and implement the lifecycle model. It's just such a sensible and powerful model for financial planning, and it's great that more people are discovering its advantages!
Total Portfolio Allocation and Withdrawal (TPAW)
User avatar
Topic Author
Ben Mathew
Posts: 2819
Joined: Tue Mar 13, 2018 11:41 am
Location: Seattle

Re: Total Portfolio Allocation and Withdrawal (TPAW)

Post by Ben Mathew »

jocdoc wrote: Mon Jul 08, 2024 7:20 am I tried to login to the planner using the email link from another computer and received the following error message.

Error
Device changed. Please open the link on the same device.

I have never received this message before. I use five different computers and several tablets throughout the week and would like to login from any device.
ScubaHogg wrote: Mon Jul 08, 2024 8:02 am You can, but when you click the link in the email it needs to automatically go to the same browser tab that sent you the link

If you use private browsing like me that won’t work. So instead I right-click the link in the email, then paste in the same browser tab that sent me the link
jocdoc wrote: Mon Jul 08, 2024 8:19 am I used outlook which opens all links in Edge (nonprivate). I requested my login Email from Edge and got the above message. I tried several times.
I just tried it now and it opened up my plan per usual without using ScubaHogg's method. i will use his method when using other browsers. Good to know. Server error initially?
You can log in on as many devices as you like. The "device changed" message is intended to mean that the device changed between where you requested the login and where the link was opened (and not that the device changed since the last time you used the planner).

As ScubaHogg mentioned, there are some additional constraints beyond opening the link on the same device. To clarify, the link does not have to open in the same browser tab, but in the same profile on the same browser on the same device that you requested the login. If that does not happen automatically, instead of clicking the link you can copy and paste it into the right profile and browser. Incognito and private browsing mode effectively create a temporary profile, so copy/pasting the link as above is always needed in those modes.

It seems like the link was opening in the same browser on the same device and everything was in non-private mode, but you still had trouble. Can you clarify if it was opening in the same profile on the browser?

Overall, these requirements are not intuitive and a source of friction, so we are planning to switch to an authentication method that is smoother.
Total Portfolio Allocation and Withdrawal (TPAW)
User avatar
Topic Author
Ben Mathew
Posts: 2819
Joined: Tue Mar 13, 2018 11:41 am
Location: Seattle

Re: Total Portfolio Allocation and Withdrawal (TPAW)

Post by Ben Mathew »

Circle the Wagons wrote: Mon Jul 08, 2024 7:17 pm
Raspberry-503 wrote: Mon Jul 08, 2024 7:08 pm
ConstantChrysalis wrote: Mon Jul 08, 2024 6:58 pm I often wonder if the specific metric here just doesn't matter much in the big picture, and the real value is in the disciplined and mechanical rebalancing that helps one avoid behavioral errors and generally encourages buying low and selling high. Maybe, similar to having a very wide range of static asset allocations that work approximately equally well for most people (like 30/70 to 70/30) with disciplined rebalancing yet worrying about dialing it into the nearest 1%.
I'm thinking along the same lines (and the VPW embraces that). I think a way to approximate that is to set SPAW and set your AA to different values and see how much it changes the montecarlo simulation.
To play devil's advocate here, one of the (many) reasons I really like TPAW is it can help the investor objectively process implications of a fluctuating ERP, even and especially in extreme situations that might push outside of an assumed "reasonable range" like 30-70% equity. But the ER assumptions chosen matter a lot -- in my case, anywhere from 27-100% suggested stocks just from the first three choices. This is why I raised the topic and appreciate hearing how folks think about it.
Yes, the equity risk premium will generally have a significant impact on the plan through the asset allocation, and also through withdrawal rates.

I personally am fairly comfortable with the default neutral "Regression Prediction" estimate. The "Conservative Estimate" seems low (as intended), especially in light of ex-US valuations. I'm uncomfortable with the raw 1/CAPE and historical estimates.
Total Portfolio Allocation and Withdrawal (TPAW)
jocdoc
Posts: 296
Joined: Wed Oct 30, 2013 5:29 am

Re: Total Portfolio Allocation and Withdrawal (TPAW)

Post by jocdoc »

Ben: I used the same pc and edge browser as I usually do. and yesterday got multiple error messages when I clicked to open the link in outlook. Same profile in Edge. I never had these error messages in the past. It did work a short time later on the same pc, browser and outlook. It worked fine this am.
ScubaHogg
Posts: 4066
Joined: Sun Nov 06, 2011 2:02 pm

Re: Total Portfolio Allocation and Withdrawal (TPAW)

Post by ScubaHogg »

Ben Mathew wrote: Tue Jul 09, 2024 12:28 am
ConstantChrysalis wrote: Mon Jul 08, 2024 6:46 am After spending some time on this, I’ve come away with an even greater appreciation for what the Matthew brothers have done for us.
Thanks for the kind words. It's exciting to be able to contribute towards helping popularize and implement the lifecycle model. It's just such a sensible and powerful model for financial planning, and it's great that more people are discovering its advantages!
It really is very cool
“Anyone may arrange his affairs so that his taxes shall be as low as possible; he is not bound to choose that pattern which best pays the treasury.” | ― Judge Learned Hand
KarenC
Posts: 270
Joined: Mon Apr 27, 2015 7:25 am

Re: Total Portfolio Allocation and Withdrawal (TPAW)

Post by KarenC »

Raspberry-503 wrote: Mon Jul 08, 2024 6:32 pm
Circle the Wagons wrote: Mon Jul 08, 2024 6:12 pm The pure 1/CAPE in TPAW is indeed 2.7%.
Yes, my bad, I read the wrong value earlier, so yes, it really is "1 divided by CAPE", I edited my post above.

I also missed that TAPW has a pop up explanation about those calculations in the Expected Returns selection section, so it tells you right there.

International CAPE is much lower, but I don't known that 1/CAPE applies there (I don't see why it should be different?), I'm not sure where to find numbers but I remember seeing 10-12 for Emerging and 20-25 for Developed.
I'm not sure if they are the best such numbers, but I use the Research Affiliates website for this. (Select the "Equity Market CAPE (boxplot)" view, and then hover over various items to get current CAPE values.)
"The first principle is that you must not fool yourself—and you are the easiest person to fool." — Richard P. Feynman
User avatar
Topic Author
Ben Mathew
Posts: 2819
Joined: Tue Mar 13, 2018 11:41 am
Location: Seattle

Re: Total Portfolio Allocation and Withdrawal (TPAW)

Post by Ben Mathew »

jocdoc wrote: Tue Jul 09, 2024 5:53 am Ben: I used the same pc and edge browser as I usually do. and yesterday got multiple error messages when I clicked to open the link in outlook. Same profile in Edge. I never had these error messages in the past. It did work a short time later on the same pc, browser and outlook. It worked fine this am.
Thanks for clarifying. That is odd. It might have been a transient issue with the authentication provider we are using. Please let me know if it happens again.
Total Portfolio Allocation and Withdrawal (TPAW)
jmk
Posts: 659
Joined: Tue Nov 01, 2011 7:48 pm

Re: Total Portfolio Allocation and Withdrawal (TPAW)

Post by jmk »

Mathew, curious why the "Decrease Risk Tolerance With Age" settings don't allow negative as well as positive values?
While the general consensus is that overall risk tolerance decreases with age, things might be a bit different in the tpaw context, where decreasing future income is already taken into account inherently, and essentials are, by definition, covered. So we're essentially talking about one's risk tolerance about general discretionary wealth (fun money). And this might go either direction.

Using myself as an example, I just retired, so my overall risk tolerance is way down than even a year ago given my inability to make up shortfalls. However, my risk tolerance for the discretionary portion of my future income hasn't changed, and I don't see it changing unless I get new information about longevity.
Post Reply