Stock/Bond Mix in Retirement: The Surprise In The Data

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
Iorek
Posts: 1569
Joined: Fri Mar 08, 2013 8:38 am

Re: Stock/Bond Mix in Retirement: The Surprise In The Data

Post by Iorek »

Personally I think “age in bonds” is primarily used as a rule of thimb for accumulators over their working life— so someone in their 20s has a very heavy stock allocation and someone approaching retirement has a substantial bond allocation.

I don’t think it was ever primarily intended— or at least I wouldn’t use it— for asset allocation in retirement.
Da5id
Posts: 5066
Joined: Fri Feb 26, 2016 7:20 am

Re: Stock/Bond Mix in Retirement: The Surprise In The Data

Post by Da5id »

martincmartin wrote: Fri Oct 22, 2021 2:11 pm This analysis really understands what a drop bigger than 50% does to your retirement. It includes the great depression, where stocks dropped by 89%.

A 67/33 stock/bond mix allows you to have a 3.9% withdrawal rate, without running out of money, whether the great depression happens at the start, middle or end of your 30 year retirement.
Bold/italics added by me above. This is the kind of statement about the SWR dataset that I find rather maddening. The claim that another catastrophic (for the country/world) scenario akin to something that has happened exactly once in the dataset can't/won't have a worse outcome is pretty weak. We supposedly have better fiscal controls and regulations these days than in the 20s/30s. If we get an 89% drawdown of stocks maybe it will go terribly worse than the great depression, who knows? And the behavioral problems of a retiree with very limited earnings potential facing a massive drawdown in an economy where nobody can get jobs would be likewise huge. How sure would you be that they'd "stay the course" and "rebalance back into stocks" because the SWR dataset says "all will be well"?

n.b. this isn't an argument against 3-4% SWR as a starting point, or a 67/33 (or 50/50) mix being generally good. I'm just not happy with the implication that the past SWR data encompasses all future outcomes, particularly when you are hypothesizing hitting something similar to one of the edge cases.
dbr
Posts: 46181
Joined: Sun Mar 04, 2007 8:50 am

Re: Stock/Bond Mix in Retirement: The Surprise In The Data

Post by dbr »

Ferdinand2014 wrote: Fri Oct 22, 2021 6:00 pm
dbr wrote: Fri Oct 22, 2021 9:12 am Quite so, the kind if analysis you are using probably can't be made simpler and still be useful. It is kind of a starting point for current approaches to need to take risk. But this is not new, since the early '90s anyway. Aside from the fact that age in bonds is just a vague shorthand for being more aggressive when young and more conservative when older, ideas that themselves do not necessarily fit everyone, the actual analysis you report suggests that there is data that supports something different from age in bonds if the criterion for risk is taking income from a portfolio and risking running out of money while still alive.

It might be interesting to examine the fate of someone with significant income from SS applying the rule age in bonds/SS as a bond which was actually Mr. Bogle's version of the rule, not that that rule is necessarily based in analysis either. At least more actual money would be placed in stocks in that case.

Some pieces missing from your analysis include at least the following considerations:

1. Some people have a lot of money and don't need to optimize SWR.

2. Those people may be motivated to maximize terminal wealth, which dictates even more stock heavy allocations.

3. A person may decide they want the minimum experience of portfolio volatility, meaning they would take the least risk necessary. That might not be the actual optimum for a simulation, but it might still be more aggressive than age in bonds, or it could even be more conservative than age in bonds.
If someones SS likely will represent 25% of estimated expenses in retirement, is it reasonable to say that if you are 100% equities with your investments, you are really 75/25 at the retirement age that you start taking SS?
No, it isn't useful to say things are something they aren't. The presence of the SS can make a big difference in what one would choose for an asset allocation, but it would be from a consideration of need, ability, and willingness to take risk and not because SS is something it isn't. I probably should not have gone down that Bogle sidetrack.
User avatar
jeffyscott
Posts: 13487
Joined: Tue Feb 27, 2007 8:12 am

Re: Stock/Bond Mix in Retirement: The Surprise In The Data

Post by jeffyscott »

Iorek wrote: Sat Oct 23, 2021 7:44 am Personally I think “age in bonds” is primarily used as a rule of thimb for accumulators over their working life— so someone in their 20s has a very heavy stock allocation and someone approaching retirement has a substantial bond allocation.

I don’t think it was ever primarily intended— or at least I wouldn’t use it— for asset allocation in retirement.
During accumulation, it effectively means reducing stock allocation as total assets increase. If you look at the dollars at risk in equities, that would still be increasing over time. Having even 100% in stocks when you have accumulated $100,000 puts far less at risk than having 50% in stocks does when you have accumulated $1,000,000 (inflation adjusted), 20 or 30 years later. The stock allocation may be 1/2 on a percentage basis, but it is 5X on a dollar basis.

If you are retired and spending 3-4% from your portfolio each year, then you can be reducing the amount in stocks as your total assets are reduced, even if the percentage remains fixed.
User avatar
siamond
Posts: 6010
Joined: Mon May 28, 2012 5:50 am

Re: Stock/Bond Mix in Retirement: The Surprise In The Data

Post by siamond »

tomsense76 wrote: Fri Oct 22, 2021 9:36 pm[...] It's much easier to think of (and fairly be afraid of) a portfolio being cut in half do to a massive decline. Though it is much harder for people to conceptualize/visualize the vicious effects of high inflation like that of the 70s. This in spite of the fact we know the latter is far more dangerous to a portfolio. Especially for a retiree who no longer is earning a paycheck (though they are getting some inflation adjustment via SS that may not cover all of their spending).
Yes, very true. Purchasing power over their entire retirement period is really what should matter the most to retirees IMHO. And that's a real (inflation-adjusted) quantity. The 1965/1966 retirees got in severe troubles not only due to the stock crisis in the early 70s, but ALSO because of the bond/inflation crisis which followed. Displaying drawdowns in real (inflation-adjusted) terms is very enlightening. "Bonds are safe"... Yeah, whatever.

Such severe inflation-adjusted drawdowns have in turn a big impact on future withdrawals, whether one uses a really dumb method (e.g. flat SWR-like approach) or a more sensible variable method (e.g. VPW/ABW, Guyton-Klinger, etc). Variable methods mitigate the problem much better, but fact is purchasing power dissipates a good deal under such trying circumstances.

Image

And if you want a true heart attack, check the longer history. Yes, you got that right. Bonds didn't recover from 40+ years... And this is the US. I have the data for the UK and it is even worse. Sure, we have TIPS nowadays and they didn't exist at that time, but those remain government-controlled instruments and call me skeptical that in time of duress (e.g. war or else), things will go accordingly to plan...

Image
HenryG
Posts: 129
Joined: Wed Aug 12, 2020 12:50 pm

Re: Stock/Bond Mix in Retirement: The Surprise In The Data

Post by HenryG »

This reminds me of a bond tent strategy, which makes sense to manage sequence of return risks in early retirement years:
https://www.kitces.com/blog/managing-po ... -red-zone/

That is generally how I'm approaching bond holdings, which means I currently look at bond holdings more as a dollar amount relative to my spending, income security, and retirement date; and less about my age and/or top-down asset allocation.
heyyou
Posts: 4461
Joined: Tue Feb 20, 2007 3:58 pm

Re: Stock/Bond Mix in Retirement: The Surprise In The Data

Post by heyyou »

My expectation is to steadily adjust future spending to match the expected future of each current situation, by spending a longevity based % amount of each recent portfolio value. Thus I steadily adjust to what is, instead of seeking the data-based optimal of a small sample of 30 year periods, in an attempt to spend more. Note that the 4% SWR often has significant money left over for all retirees who do not experience the historical worst case period for that particular method, yet it would have failed for some retirees who had not finished their 30 years when the method was being researched. Yes, the improvements are now known, but I just do not need to boost my income for every single tenth a percent of inflation while ignoring my remaining portfolio balance, for thirty years. Both under spending and over spending are risks, so steadily adapt your spending as your specific returns vary, during your retirement years.
User avatar
willthrill81
Posts: 32250
Joined: Thu Jan 26, 2017 2:17 pm
Location: USA
Contact:

Re: Stock/Bond Mix in Retirement: The Surprise In The Data

Post by willthrill81 »

HenryG wrote: Sat Oct 23, 2021 11:17 am This reminds me of a bond tent strategy, which makes sense to manage sequence of return risks in early retirement years:
Actually, it doesn't. There have been several threads debunking the 'bond tent' approach.

It's actually pretty simple to see how the strategy could lead to a worse outcome. If stocks do well when you're initially tilted toward bonds but then stocks do poorly when you're later on tilted toward them, you'll come out behind a fixed AA.
The Sensible Steward
User avatar
Ben Mathew
Posts: 2720
Joined: Tue Mar 13, 2018 11:41 am
Location: Seattle

Re: Stock/Bond Mix in Retirement: The Surprise In The Data

Post by Ben Mathew »

heyyou wrote: Sat Oct 23, 2021 12:58 pm My expectation is to steadily adjust future spending to match the expected future of each current situation, by spending a longevity based % amount of each recent portfolio value. Thus I steadily adjust to what is, instead of seeking the data-based optimal of a small sample of 30 year periods, in an attempt to spend more. Note that the 4% SWR often has significant money left over for all retirees who do not experience the historical worst case period for that particular method, yet it would have failed for some retirees who had not finished their 30 years when the method was being researched. Yes, the improvements are now known, but I just do not need to boost my income for every single tenth a percent of inflation while ignoring my remaining portfolio balance, for thirty years. Both under spending and over spending are risks, so steadily adapt your spending as your specific returns vary, during your retirement years.
This makes a lot of sense.

Retirement planning would be well served by adopting variable withdrawal planning concepts. It's hard to admit that you don't know exactly what your withdrawal will be at age 80. But that's the reality of consuming out of a risky portfolio, and it's best to embrace it and plan around it.
Total Portfolio Allocation and Withdrawal (TPAW)
gurusw
Posts: 310
Joined: Sun May 11, 2008 9:37 am

Re: Stock/Bond Mix in Retirement: The Surprise In The Data

Post by gurusw »

How do I apply these learnings to a 529 plan? Maybe I should not go below 50-50 mix in 529?
Northern Flicker
Posts: 15369
Joined: Fri Apr 10, 2015 12:29 am

Re: Stock/Bond Mix in Retirement: The Surprise In The Data

Post by Northern Flicker »

Ferdinand2014 wrote: If someones SS likely will represent 25% of estimated expenses in retirement, is it reasonable to say that if you are 100% equities with your investments, you are really 75/25 at the retirement age that you start taking SS?
While you theoretically could model it that way, it is easier and less error prone to view the asset allocation for what it is, and model the expenses you need to cover with retirement savings as projected residual expenses not covered by SS.
Iorek
Posts: 1569
Joined: Fri Mar 08, 2013 8:38 am

Re: Stock/Bond Mix in Retirement: The Surprise In The Data

Post by Iorek »

gurusw wrote: Sat Oct 23, 2021 3:36 pm How do I apply these learnings to a 529 plan? Maybe I should not go below 50-50 mix in 529?
I would not apply any of this to a 529 plan. A 529 plan involves investing over a period of 0-18 years with the goal of (generally) paying it all out over 4 years. Investing for retirement means investing for 30+ years with the goal of paying very small amounts each year.

I’m not sure what you mean by not going “below” 50/50 but you want to be careful about ending up in a situation where you can’t fund a given year’s expenses because of a drop in market value of equities. What that looks like for a 529 is very different from what it looks like for retirement.
dbr
Posts: 46181
Joined: Sun Mar 04, 2007 8:50 am

Re: Stock/Bond Mix in Retirement: The Surprise In The Data

Post by dbr »

gurusw wrote: Sat Oct 23, 2021 3:36 pm How do I apply these learnings to a 529 plan? Maybe I should not go below 50-50 mix in 529?
More likely in the years you are actually spending the money you should have none of the assets in equities, and before that as well.

As another post points out spending the money in a short time and high rate is completely different from a long time at a low rate.
User avatar
patrick013
Posts: 3301
Joined: Mon Jul 13, 2015 7:49 pm

Re: Stock/Bond Mix in Retirement: The Surprise In The Data

Post by patrick013 »

martincmartin wrote: Fri Oct 22, 2021 8:29 am
The problem with "your age in bonds."

But some say "age - 20," others say constant allocation. Let data decide! For a 4% withdrawal rate, a constant 50/50 allocation has a 91.67% success rate on cFIREsim. How much better is "age in bonds"?
I've always thought "Age-in-bonds Stop at 50" was conservative but not that
much. Is there a way to backtest that ?
age in bonds, buy-and-hold, 10 year business cycle
User avatar
Topic Author
martincmartin
Posts: 900
Joined: Wed Jul 02, 2014 3:04 pm
Location: Boston, MA USA

Re: Stock/Bond Mix in Retirement: The Surprise In The Data

Post by martincmartin »

patrick013 wrote: Sun Oct 24, 2021 12:07 pm
martincmartin wrote: Fri Oct 22, 2021 8:29 am
The problem with "your age in bonds."

But some say "age - 20," others say constant allocation. Let data decide! For a 4% withdrawal rate, a constant 50/50 allocation has a 91.67% success rate on cFIREsim. How much better is "age in bonds"?
I've always thought "Age-in-bonds Stop at 50" was conservative but not that
much. Is there a way to backtest that ?
If your age in retirement ranges from 65 to 95, then your rule turns into the 50/50 allocation above.
HenryG
Posts: 129
Joined: Wed Aug 12, 2020 12:50 pm

Re: Stock/Bond Mix in Retirement: The Surprise In The Data

Post by HenryG »

willthrill81 wrote: Sat Oct 23, 2021 1:05 pm
HenryG wrote: Sat Oct 23, 2021 11:17 am This reminds me of a bond tent strategy, which makes sense to manage sequence of return risks in early retirement years:
Actually, it doesn't. There have been several threads debunking the 'bond tent' approach.

It's actually pretty simple to see how the strategy could lead to a worse outcome. If stocks do well when you're initially tilted toward bonds but then stocks do poorly when you're later on tilted toward them, you'll come out behind a fixed AA.
I don't understand what you're saying about tilting initially and later. Can you elaborate?
User avatar
willthrill81
Posts: 32250
Joined: Thu Jan 26, 2017 2:17 pm
Location: USA
Contact:

Re: Stock/Bond Mix in Retirement: The Surprise In The Data

Post by willthrill81 »

HenryG wrote: Sun Oct 24, 2021 1:11 pm
willthrill81 wrote: Sat Oct 23, 2021 1:05 pm
HenryG wrote: Sat Oct 23, 2021 11:17 am This reminds me of a bond tent strategy, which makes sense to manage sequence of return risks in early retirement years:
Actually, it doesn't. There have been several threads debunking the 'bond tent' approach.

It's actually pretty simple to see how the strategy could lead to a worse outcome. If stocks do well when you're initially tilted toward bonds but then stocks do poorly when you're later on tilted toward them, you'll come out behind a fixed AA.
I don't understand what you're saying about tilting initially and later. Can you elaborate?
That's what the bond tent strategy does. Early in your retirement, you have a comparatively high allocation to bonds and then progressively increase your stock allocation as you age.

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

The graph should help to illustrate my point above. If stocks do well when you're mostly allocated toward bonds but then do poorly when you're mostly allocated toward stocks, you will have a worse outcome than having had a fixed asset allocation all along.
The Sensible Steward
HenryG
Posts: 129
Joined: Wed Aug 12, 2020 12:50 pm

Re: Stock/Bond Mix in Retirement: The Surprise In The Data

Post by HenryG »

willthrill81 wrote: Sun Oct 24, 2021 1:16 pm If stocks do well when you're mostly allocated toward bonds but then do poorly when you're mostly allocated toward stocks, you will have a worse outcome than having had a fixed asset allocation all along.
Of course, but wouldn't you do better on the inverse? You're more protected if stocks do poorly early, and you can tolerate a higher stock allocation later because you've made it through a high period of risk (i.e., early retirement years). That's one of the benefits I see in the tent strategy; maintaining a higher bond allocation when it matters most. It sounds like you're recommending a fixed AA throughout life which doesn't sound practical to me (maybe I'm misinterpreting).
User avatar
willthrill81
Posts: 32250
Joined: Thu Jan 26, 2017 2:17 pm
Location: USA
Contact:

Re: Stock/Bond Mix in Retirement: The Surprise In The Data

Post by willthrill81 »

HenryG wrote: Sun Oct 24, 2021 1:34 pm
willthrill81 wrote: Sun Oct 24, 2021 1:16 pm If stocks do well when you're mostly allocated toward bonds but then do poorly when you're mostly allocated toward stocks, you will have a worse outcome than having had a fixed asset allocation all along.
Of course, but wouldn't you do better on the inverse? You're more protected if stocks do poorly early, and you can tolerate a higher stock allocation later because you've made it through a high period of risk (i.e., early retirement years). That's one of the benefits I see in the tent strategy; maintaining a higher bond allocation when it matters most. It sounds like you're recommending a fixed AA throughout life which doesn't sound practical to me (maybe I'm misinterpreting).
It's true that if stocks do poorly when you're tilted toward bonds, but if the inverse happens, this strategy will come out behind a fixed AA. In general, we should avoid strategies that only pay off if one-sided events occur. Others have done subsequent testing to the original bond tent approach as put forth by Kitces and Pfau and found that it hurt retirees at least as often as it helped them.

Issues with the bond tent approach have been discussed at length in other threads, such as this one, this one, this one, and this one.

I don't see why maintaining a steady AA throughout retirement doesn't sound practical. It actually sounds much more practical than increasing one's allocation to stocks as one ages. The general trend among retirees seems to be that their tolerance for volatility goes down with age. Also, there is no need to try to figure out how much one's stock allocation should change over time nor when; you just maintain a steady AA.
Last edited by willthrill81 on Sun Oct 24, 2021 1:56 pm, edited 1 time in total.
The Sensible Steward
Marseille07
Posts: 16054
Joined: Fri Nov 06, 2020 12:41 pm

Re: Stock/Bond Mix in Retirement: The Surprise In The Data

Post by Marseille07 »

willthrill81 wrote: Sun Oct 24, 2021 1:53 pm I don't see why maintaining a steady AA throughout retirement doesn't sound practical. It actually sounds much more practical than increasing one's allocation to stocks as one ages. The general trend among retirees seems to be that their tolerance for volatility goes down with age. Also, there is no need to try to figure out how much one's stock allocation should change over time nor when; you just maintain a steady AA.
I'm with you, but I also understand the notion of slowing down (i.e. lowering AA) once you get to something sizable like 10M.
User avatar
Ben Mathew
Posts: 2720
Joined: Tue Mar 13, 2018 11:41 am
Location: Seattle

Re: Stock/Bond Mix in Retirement: The Surprise In The Data

Post by Ben Mathew »

HenryG wrote: Sun Oct 24, 2021 1:34 pm
willthrill81 wrote: Sun Oct 24, 2021 1:16 pm If stocks do well when you're mostly allocated toward bonds but then do poorly when you're mostly allocated toward stocks, you will have a worse outcome than having had a fixed asset allocation all along.
Of course, but wouldn't you do better on the inverse? You're more protected if stocks do poorly early, and you can tolerate a higher stock allocation later because you've made it through a high period of risk (i.e., early retirement years). That's one of the benefits I see in the tent strategy; maintaining a higher bond allocation when it matters most. It sounds like you're recommending a fixed AA throughout life which doesn't sound practical to me (maybe I'm misinterpreting).
The glidepath that minimizes sequence of return risk is downward sloping prior to retirement and fixed in retirement (putting aside pensions). The downward slope prior to retirement is to adjust for the fact that not all contributions are in. Once retirement starts, there's no more future contributions, so the asset allocation becomes fixed.

The fixed allocation during retirement may seem surprising when you consider that the age 65 portfolio is much larger than the portfolio at ages 75 or 85 will be. So how can the same stock percentage result in constant risk?

To see why, look at it like this: At age 65, the retirement account holds money that will fund many different ages: age 66, 67,...,100. At age 85, it will hold money that will fund only ages 86, 87, ..., 100.

By holding the same fixed percentage in retirement, you are simply applying the same allocation to each age bucket every year. A 10% drop in stocks at age 65 affects future consumption by the same amount as a 10% drop in stocks at age 85 would do. You are not overexposed at age 65 relative to other ages. There are just more age buckets you're funding. Age 65 returns does matter more because it impacts more ages. But each age that is impacted is impacted the same.

If instead you increase stock allocation over time (bond bridge, rising equity etc.) then the age 85 stock return will impact future ages more than the age 65 stock return. Age 90 consumption will be more dependent on age 85 return than on age 65 return. Risk is not being kept constant leading to higher overall risk for the same expected return.
Last edited by Ben Mathew on Sun Oct 24, 2021 2:11 pm, edited 1 time in total.
Total Portfolio Allocation and Withdrawal (TPAW)
MIretired
Posts: 1059
Joined: Fri Sep 06, 2013 12:35 pm
Location: Midwest, US

Re: Stock/Bond Mix in Retirement: The Surprise In The Data

Post by MIretired »

Marseille07 wrote: Sun Oct 24, 2021 1:55 pm
willthrill81 wrote: Sun Oct 24, 2021 1:53 pm I don't see why maintaining a steady AA throughout retirement doesn't sound practical. It actually sounds much more practical than increasing one's allocation to stocks as one ages. The general trend among retirees seems to be that their tolerance for volatility goes down with age. Also, there is no need to try to figure out how much one's stock allocation should change over time nor when; you just maintain a steady AA.
I'm with you, but I also understand the notion of slowing down (i.e. lowering AA) once you get to something sizable like 10M.
I agree that near or after retirement, it's more about real future dollars needed and not about %'s strategies.
User avatar
willthrill81
Posts: 32250
Joined: Thu Jan 26, 2017 2:17 pm
Location: USA
Contact:

Re: Stock/Bond Mix in Retirement: The Surprise In The Data

Post by willthrill81 »

Marseille07 wrote: Sun Oct 24, 2021 1:55 pm
willthrill81 wrote: Sun Oct 24, 2021 1:53 pm I don't see why maintaining a steady AA throughout retirement doesn't sound practical. It actually sounds much more practical than increasing one's allocation to stocks as one ages. The general trend among retirees seems to be that their tolerance for volatility goes down with age. Also, there is no need to try to figure out how much one's stock allocation should change over time nor when; you just maintain a steady AA.
I'm with you, but I also understand the notion of slowing down (i.e. lowering AA) once you get to something sizable like 10M.
That's a different animal entirely though. The thinking behind that approach is that once you reach a certain level of wealth, maintaining that wealth is more important than increasing it, hence, dialing down stock exposure, though not too much. Conversely, the bond tent approach reduces stock exposure and then increases it later.

On a different note though, one could just as easily argue that those whose portfolios are very large in relation to their planned withdrawals have a greater ability to take on risk. They could justly leave their stock allocation as is or even increase it, but they could also reduce their stock exposure. This was explained in this post earlier this year.
The Sensible Steward
Marseille07
Posts: 16054
Joined: Fri Nov 06, 2020 12:41 pm

Re: Stock/Bond Mix in Retirement: The Surprise In The Data

Post by Marseille07 »

willthrill81 wrote: Sun Oct 24, 2021 2:00 pm On a different note though, one could just as easily argue that those whose portfolios are very large in relation to their planned withdrawals have a greater ability to take on risk. They could justly leave their stock allocation as is or even increase it, but they could also reduce their stock exposure. This was explained in this post earlier this year.
Yes, this is my plan actually. The need for fixed income remains more or less constant, as long as that's secured we can invest the rest.
HenryG
Posts: 129
Joined: Wed Aug 12, 2020 12:50 pm

Re: Stock/Bond Mix in Retirement: The Surprise In The Data

Post by HenryG »

willthrill81 wrote: Sun Oct 24, 2021 1:53 pm In general, we should avoid strategies that only pay off if one-sided events occur.
Running out of money is more damaging than having excess money IMO.
willthrill81 wrote: Sun Oct 24, 2021 1:53 pm Issues with the bond tent approach have been discussed at length in other threads, such as this one, this one, this one, and this one.
There is support for a bond tent strategy in the threads you posted. The arguments against seem to be 1) just adjust your withdrawal percentage down, or 2) just go back to work. I agree, those can be pursued when any investment strategy begins to fail.
willthrill81 wrote: Sun Oct 24, 2021 1:53 pm I don't see why maintaining a steady AA throughout retirement doesn't sound practical. It actually sounds much more practical than increasing one's allocation to stocks as one ages.
Why maintain a steady AA in retirement if you don't need the stability of the bond holdings after a certain point? I.e., if stock grows, why sell it to buy bonds just to maintain a fixed AA percentage, if your bond holdings in $$s already provide sufficient security relative to your spending? Just hold your bonds, and let your stocks run up (or down) indefinitely.

It seems we both agree that increasing AA to bonds at or near retirement age is prudent (pls correct me if I misread). I'm not advocating increasing AA to stocks in late retirement because you need the higher expected returns; I'm advocating it because at some point you don't need the returns...you're investing a perpetual portfolio geared for your heirs. Greater bond holdings for security when you need it (early retirement, the upside to the tent so to speak); and greater stock holdings when you don't (later in life the downside to the tent so to speak). What am I missing, I'm sure it's something? Does this only make sense under certain conditions?
HenryG
Posts: 129
Joined: Wed Aug 12, 2020 12:50 pm

Re: Stock/Bond Mix in Retirement: The Surprise In The Data

Post by HenryG »

Ben Mathew wrote: Sun Oct 24, 2021 1:59 pm A 10% drop in stocks at age 65 affects future consumption by the same amount as a 10% drop in stocks at age 85 would do.
I think the affect would be different, because annual spending expectations and potential drawdown periods (life expectancy) are likely to differ at those ages. Portfolio size would matter as well.
User avatar
Ben Mathew
Posts: 2720
Joined: Tue Mar 13, 2018 11:41 am
Location: Seattle

Re: Stock/Bond Mix in Retirement: The Surprise In The Data

Post by Ben Mathew »

HenryG wrote: Sun Oct 24, 2021 9:05 pm
Ben Mathew wrote: Sun Oct 24, 2021 1:59 pm A 10% drop in stocks at age 65 affects future consumption by the same amount as a 10% drop in stocks at age 85 would do.
I think the affect would be different, because annual spending expectations and potential drawdown periods (life expectancy) are likely to differ at those ages. Portfolio size would matter as well.
Below is a simulation that shows how a fixed asset allocation combined with variable withdrawals makes retirement spending at any age equally sensitive to stock returns at prior ages. It doesn't model longevity risk, but does include variable spending expectations and portfolio size.

Time diversification of stock risk
Total Portfolio Allocation and Withdrawal (TPAW)
User avatar
willthrill81
Posts: 32250
Joined: Thu Jan 26, 2017 2:17 pm
Location: USA
Contact:

Re: Stock/Bond Mix in Retirement: The Surprise In The Data

Post by willthrill81 »

HenryG wrote: Sun Oct 24, 2021 8:49 pm
willthrill81 wrote: Sun Oct 24, 2021 1:53 pm In general, we should avoid strategies that only pay off if one-sided events occur.
Running out of money is more damaging than having excess money IMO.
willthrill81 wrote: Sun Oct 24, 2021 1:53 pm Issues with the bond tent approach have been discussed at length in other threads, such as this one, this one, this one, and this one.
There is support for a bond tent strategy in the threads you posted. The arguments against seem to be 1) just adjust your withdrawal percentage down, or 2) just go back to work. I agree, those can be pursued when any investment strategy begins to fail.
willthrill81 wrote: Sun Oct 24, 2021 1:53 pm I don't see why maintaining a steady AA throughout retirement doesn't sound practical. It actually sounds much more practical than increasing one's allocation to stocks as one ages.
Why maintain a steady AA in retirement if you don't need the stability of the bond holdings after a certain point? I.e., if stock grows, why sell it to buy bonds just to maintain a fixed AA percentage, if your bond holdings in $$s already provide sufficient security relative to your spending? Just hold your bonds, and let your stocks run up (or down) indefinitely.

It seems we both agree that increasing AA to bonds at or near retirement age is prudent (pls correct me if I misread). I'm not advocating increasing AA to stocks in late retirement because you need the higher expected returns; I'm advocating it because at some point you don't need the returns...you're investing a perpetual portfolio geared for your heirs. Greater bond holdings for security when you need it (early retirement, the upside to the tent so to speak); and greater stock holdings when you don't (later in life the downside to the tent so to speak). What am I missing, I'm sure it's something? Does this only make sense under certain conditions?
Ben Mathew explained part of the mathematical problem with increasing one's stock exposure later in retirement.

Again, if stocks do well when you're heavily tilted to bonds early in retirement, you'll come out behind with a bond tent strategy. And if bonds do poorly during that decade, you're doubly hosed.
The Sensible Steward
secondopinion
Posts: 6011
Joined: Wed Dec 02, 2020 12:18 pm

Re: Stock/Bond Mix in Retirement: The Surprise In The Data

Post by secondopinion »

willthrill81 wrote: Sun Oct 24, 2021 9:27 pm
HenryG wrote: Sun Oct 24, 2021 8:49 pm
willthrill81 wrote: Sun Oct 24, 2021 1:53 pm In general, we should avoid strategies that only pay off if one-sided events occur.
Running out of money is more damaging than having excess money IMO.
willthrill81 wrote: Sun Oct 24, 2021 1:53 pm Issues with the bond tent approach have been discussed at length in other threads, such as this one, this one, this one, and this one.
There is support for a bond tent strategy in the threads you posted. The arguments against seem to be 1) just adjust your withdrawal percentage down, or 2) just go back to work. I agree, those can be pursued when any investment strategy begins to fail.
willthrill81 wrote: Sun Oct 24, 2021 1:53 pm I don't see why maintaining a steady AA throughout retirement doesn't sound practical. It actually sounds much more practical than increasing one's allocation to stocks as one ages.
Why maintain a steady AA in retirement if you don't need the stability of the bond holdings after a certain point? I.e., if stock grows, why sell it to buy bonds just to maintain a fixed AA percentage, if your bond holdings in $$s already provide sufficient security relative to your spending? Just hold your bonds, and let your stocks run up (or down) indefinitely.

It seems we both agree that increasing AA to bonds at or near retirement age is prudent (pls correct me if I misread). I'm not advocating increasing AA to stocks in late retirement because you need the higher expected returns; I'm advocating it because at some point you don't need the returns...you're investing a perpetual portfolio geared for your heirs. Greater bond holdings for security when you need it (early retirement, the upside to the tent so to speak); and greater stock holdings when you don't (later in life the downside to the tent so to speak). What am I missing, I'm sure it's something? Does this only make sense under certain conditions?

Ben Mathew explained part of the mathematical problem with increasing one's stock exposure later in retirement.


Again, if stocks do well when you're heavily tilted to bonds early in retirement, you'll come out behind with a bond tent strategy. And if bonds do poorly during that decade, you're doubly hosed.
Of course, if this is excess to the funds need for retirement, heirs might benefit from the aggressiveness.
Passive investing: not about making big bucks but making profits. Active investing: not about beating the market but meeting goals. Speculation: not about timing the market but taking profitable risks.
User avatar
Ben Mathew
Posts: 2720
Joined: Tue Mar 13, 2018 11:41 am
Location: Seattle

Re: Stock/Bond Mix in Retirement: The Surprise In The Data

Post by Ben Mathew »

secondopinion wrote: Sun Oct 24, 2021 10:12 pm
willthrill81 wrote: Sun Oct 24, 2021 9:27 pm
HenryG wrote: Sun Oct 24, 2021 8:49 pm
willthrill81 wrote: Sun Oct 24, 2021 1:53 pm In general, we should avoid strategies that only pay off if one-sided events occur.
Running out of money is more damaging than having excess money IMO.
willthrill81 wrote: Sun Oct 24, 2021 1:53 pm Issues with the bond tent approach have been discussed at length in other threads, such as this one, this one, this one, and this one.
There is support for a bond tent strategy in the threads you posted. The arguments against seem to be 1) just adjust your withdrawal percentage down, or 2) just go back to work. I agree, those can be pursued when any investment strategy begins to fail.
willthrill81 wrote: Sun Oct 24, 2021 1:53 pm I don't see why maintaining a steady AA throughout retirement doesn't sound practical. It actually sounds much more practical than increasing one's allocation to stocks as one ages.
Why maintain a steady AA in retirement if you don't need the stability of the bond holdings after a certain point? I.e., if stock grows, why sell it to buy bonds just to maintain a fixed AA percentage, if your bond holdings in $$s already provide sufficient security relative to your spending? Just hold your bonds, and let your stocks run up (or down) indefinitely.

It seems we both agree that increasing AA to bonds at or near retirement age is prudent (pls correct me if I misread). I'm not advocating increasing AA to stocks in late retirement because you need the higher expected returns; I'm advocating it because at some point you don't need the returns...you're investing a perpetual portfolio geared for your heirs. Greater bond holdings for security when you need it (early retirement, the upside to the tent so to speak); and greater stock holdings when you don't (later in life the downside to the tent so to speak). What am I missing, I'm sure it's something? Does this only make sense under certain conditions?

Ben Mathew explained part of the mathematical problem with increasing one's stock exposure later in retirement.


Again, if stocks do well when you're heavily tilted to bonds early in retirement, you'll come out behind with a bond tent strategy. And if bonds do poorly during that decade, you're doubly hosed.
Of course, if this is excess to the funds need for retirement, heirs might benefit from the aggressiveness.
If the goal is to live on what you need and leave the rest to heirs, then it makes sense to keep the inheritance at a more aggressive allocation. An 80/20 allocation on the whole portfolio may actually be something like 30/70 on retirement funds and 100/0 on the inheritance funds. Here too, it helps to maintain a fixed allocation on each subportfolio. The resulting time diversification will minimize risk (given the expected return) for both goals. The retirement funds will have minimal sequence of return risk and the inheritance funds will have no sequence of return risk.
Total Portfolio Allocation and Withdrawal (TPAW)
Northern Flicker
Posts: 15369
Joined: Fri Apr 10, 2015 12:29 am

Re: Stock/Bond Mix in Retirement: The Surprise In The Data

Post by Northern Flicker »

willthrill81 wrote: Again, if stocks do well when you're heavily tilted to bonds early in retirement, you'll come out behind with a bond tent strategy. And if bonds do poorly during that decade, you're doubly hosed.
While I am not advocating for a bond tent strategy, you are implicitly using a metric of highest return in evaluating the strategy. Sure, there are scenarios where a bond tent has a lower return. The question is, however, how does a bond tent affect the probability of running out of money?

I don't think that question has ever been answered.
HenryG
Posts: 129
Joined: Wed Aug 12, 2020 12:50 pm

Re: Stock/Bond Mix in Retirement: The Surprise In The Data

Post by HenryG »

Ben Mathew wrote: Sun Oct 24, 2021 10:46 pm If the goal is to live on what you need and leave the rest to heirs, then it makes sense to keep the inheritance at a more aggressive allocation. An 80/20 allocation on the whole portfolio may actually be something like 30/70 on retirement funds and 100/0 on the inheritance funds.
The two funds help explain this, thanks for the example. And the more total $ you can afford to have in the aggressive (inheritance funds) part of the portfolio, the higher likelihood the overall portfolio tilts to a more aggressive AA over time. Which sort of aligns with the AA example the OP mentioned:
martincmartin wrote: Fri Oct 22, 2021 8:29 am start with 35% equities at 65 years old, then *increase* to 100% equities at 95 years old.
The OP didn't say, but if someone constructed their portfolio this way, I assume they would build up the bond holdings ahead of age 65. Which is why the whole scenario reminds me of a bond tent - the prospective retiree builds a reserve of bonds in the final decade leading up to retirement, and then spends down that bond reserve in the early years of retirement itself (allowing equity exposure to return to normal) - which was my original point.

The timing of building up/drawing down bonds or the level to which they are built up in the bond tent scenario could certainly be influenced by portfolio size, expected spend, time horizon, etc. - i.e., your ability, willingness, and need to take risk. I don't think I'll ever be as conservative as 65% bonds, for example, but that's just my view today.
User avatar
Topic Author
martincmartin
Posts: 900
Joined: Wed Jul 02, 2014 3:04 pm
Location: Boston, MA USA

Re: Stock/Bond Mix in Retirement: The Surprise In The Data

Post by martincmartin »

HenryG wrote: Mon Oct 25, 2021 6:34 am
martincmartin wrote: Fri Oct 22, 2021 8:29 am start with 35% equities at 65 years old, then *increase* to 100% equities at 95 years old.
The OP didn't say, but if someone constructed their portfolio this way, I assume they would build up the bond holdings ahead of age 65. Which is why the whole scenario reminds me of a bond tent
Yes, this is where a bond tent comes from. If you do the back testing with adding funds instead of spending them, so for the accumulation phase, the optimal is a falling equity glide path. So falling during accumulation, rising during retirement is a bond tent.

Note that this has nothing to do with leaving a legacy. In this analysis, you set the SWR to whatever has you just running out of money at the end of the worst case retirement. For 30 years with S &P 500 and Total Bond Market, this is roughly 3.9%, IIRC. The point is, the optimization isn't looking at leaving a legacy, the end value of the portfolio at 95 years old is $0. Obviously, this isn't what you actually do in practice. In practice you build in a buffer. But the rising equity glide path is found by the optimizer not considering any legacy.
User avatar
CyclingDuo
Posts: 6009
Joined: Fri Jan 06, 2017 8:07 am

Re: Stock/Bond Mix in Retirement: The Surprise In The Data

Post by CyclingDuo »

Iorek wrote: Sun Oct 24, 2021 10:45 am
gurusw wrote: Sat Oct 23, 2021 3:36 pm How do I apply these learnings to a 529 plan? Maybe I should not go below 50-50 mix in 529?
I would not apply any of this to a 529 plan. A 529 plan involves investing over a period of 0-18 years with the goal of (generally) paying it all out over 4 years. Investing for retirement means investing for 30+ years with the goal of paying very small amounts each year.

I’m not sure what you mean by not going “below” 50/50 but you want to be careful about ending up in a situation where you can’t fund a given year’s expenses because of a drop in market value of equities. What that looks like for a 529 is very different from what it looks like for retirement.
Getting OT regarding SWRs and AA for a retirement portfolio, but just wanted to chime in about college education funds as we are not in the camp that would subscribe to the 0-18 year time frame being the correct fit for everyone. Some need a longer time frame.

My spouse and I both earned advanced degrees (as did our parents). So when we had children and immediately began college funding for our own children as soon as I had their SS #'s, we just assumed that they would also be pursuing advanced degrees starting 22 years later after they finished their undergraduate degrees. So we front loaded their college fund accounts.

All in all, the investing time horizon was a minimum of 25-26 years from start to finish. Some students that pursue a PhD or Doctorate should have parents focusing on a longer period of years than 0-18. Perhaps even 28-30+ years. Mix in scholarships, paid internships, paid assistantships, summer and holiday jobs to help cover costs - and the decumulation of the funds that were saved to cover the costs were spread out enough to at least suggest - based on our experience - those parents who have children that will pursue advanced degrees to think well beyond a basic 0-18 year time frame for investing.

We did it all by being invested in their accounts with an AA of 100% equities from start to finish. End result, both graduated debt free and began their working lives with the leftover amounts in the college funds (those leftover accounts are 6 figures now). That included being invested through the 2000-2009 period and the first child began college in 2011 when semester one payment was due. Things bounced back nicely following the March 2009 lows to cover all the college and graduate school funding years throughout the decade.

Yes, it is a different scenario time wise for a student only pursuing an undergraduate degree, but I thought it worth pointing out since many do go on to get advanced degrees and the time horizon in a parent's planning should consider including that - both in length as well as AA. At least in our opinion.

CyclingDuo
"Save like a pessimist, invest like an optimist." - Morgan Housel | "Pick a bushel, save a peck!" - Grandpa
User avatar
willthrill81
Posts: 32250
Joined: Thu Jan 26, 2017 2:17 pm
Location: USA
Contact:

Re: Stock/Bond Mix in Retirement: The Surprise In The Data

Post by willthrill81 »

Northern Flicker wrote: Mon Oct 25, 2021 1:56 am
willthrill81 wrote: Again, if stocks do well when you're heavily tilted to bonds early in retirement, you'll come out behind with a bond tent strategy. And if bonds do poorly during that decade, you're doubly hosed.
While I am not advocating for a bond tent strategy, you are implicitly using a metric of highest return in evaluating the strategy. Sure, there are scenarios where a bond tent has a lower return. The question is, however, how does a bond tent affect the probability of running out of money?

I don't think that question has ever been answered.
Karsten Jeske examined the bond tent approach for very early retirees (i.e., 60 year retirements) and found that there was a very small benefit, roughly a .2% increase in the SWR, from starting at a 60% stock allocation that glided up to 100% stocks. He also found, as I noted, that the strategy has come out behind a fixed AA if stocks do well at first and poorly later on. But, more importantly, the bond tent approach did not produce consistently meaningful improvement in the SWR. For instance, the 60% to 80%, 80% to 100%, 40% to 80%, 20% to 80%, and 40% to 80% glidepaths he tested only improved the SWR of the baseline 75% AA by anywhere from 2 to 16 basis points. This is actually similar to what Kitces found with the approach, namely that the historic benefit of the bond tent approach would have been tiny. As such, we must question whether these results were mere historic artifacts that are not likely to be replicated going forward. Further, I doubt that most retirees would be comfortable with a 60% stock allocation going up to 100% in their old age.

Those who think that a historic maximum increase in the SWR of 20 basis points is worthwhile should examine the historic impact of other strategies that have had a far more substantial impact on SWRs, such as factors, gold, real estate, and trend following. For instance, since 1970, the 30 year SWR for a portfolio with 30% U.S. stock, 30% ex-U.S. stock, and 40% intermediate term Treasuries was 4.6%. But with a tilt toward SCV (i.e., 20% U.S. stock / 10% U.S. SCV / 20% ex-U.S. stock / 10% ex-U.S. SCV / 40% ITT), the 30 year SWR was 5.1%, a 50 basis point increase, more than double what the bond tent approach would have achieved in the best case.
The Sensible Steward
User avatar
Ben Mathew
Posts: 2720
Joined: Tue Mar 13, 2018 11:41 am
Location: Seattle

Re: Stock/Bond Mix in Retirement: The Surprise In The Data

Post by Ben Mathew »

HenryG wrote: Mon Oct 25, 2021 6:34 am
Ben Mathew wrote: Sun Oct 24, 2021 10:46 pm If the goal is to live on what you need and leave the rest to heirs, then it makes sense to keep the inheritance at a more aggressive allocation. An 80/20 allocation on the whole portfolio may actually be something like 30/70 on retirement funds and 100/0 on the inheritance funds.
The two funds help explain this, thanks for the example. And the more total $ you can afford to have in the aggressive (inheritance funds) part of the portfolio, the higher likelihood the overall portfolio tilts to a more aggressive AA over time.
Yes, with legacy considerations the overall portfolio would get more aggressive over time because towards the end of life most of the money left is for the heirs and is being allocated aggressively.

But OP's modeling does not include legacy considerations, and so his results are not being driven by this.
Total Portfolio Allocation and Withdrawal (TPAW)
Northern Flicker
Posts: 15369
Joined: Fri Apr 10, 2015 12:29 am

Re: Stock/Bond Mix in Retirement: The Surprise In The Data

Post by Northern Flicker »

willthrill81 wrote: But, more importantly, the bond tent approach did not produce consistently meaningful improvement in the SWR. For instance, the 60% to 80%, 80% to 100%, 40% to 80%, 20% to 80%, and 40% to 80% glidepaths he tested only improved the SWR of the baseline 75% AA by anywhere from 2 to 16 basis points. This is actually similar to what Kitces found with the approach, namely that the historic benefit of the bond tent approach would have been tiny. As such, we must question whether these results were mere historic artifacts that are not likely to be replicated going forward.
Bingo. Had the work been done with a random sample of independent trials, the very modest measured increase in SWR perhaps was statistically significant, but with the sample bias of historical financial return data, it seems to be in the noise.
dbr
Posts: 46181
Joined: Sun Mar 04, 2007 8:50 am

Re: Stock/Bond Mix in Retirement: The Surprise In The Data

Post by dbr »

Northern Flicker wrote: Mon Oct 25, 2021 12:38 pm
willthrill81 wrote: But, more importantly, the bond tent approach did not produce consistently meaningful improvement in the SWR. For instance, the 60% to 80%, 80% to 100%, 40% to 80%, 20% to 80%, and 40% to 80% glidepaths he tested only improved the SWR of the baseline 75% AA by anywhere from 2 to 16 basis points. This is actually similar to what Kitces found with the approach, namely that the historic benefit of the bond tent approach would have been tiny. As such, we must question whether these results were mere historic artifacts that are not likely to be replicated going forward.
Bingo. Had the work been done with a random sample of independent trials, the very modest measured increase in SWR perhaps was statistically significant, but with the sample bias of historical financial return data, it seems to be in the noise.
I think this just indicates that it take a lot of proving to believe that messing around with asset allocation has much of an effect on the long term success or failure of a withdrawal plan, buckets, tents, whatever.

It might be more productive to seek improvements in plans that mess around with the withdrawal rate. Even then I am not sure it is that helpful.

A person can also mess around with the income stream by purchasing annuities or loading up a TIPS ladder. That too takes some proving.

At bottom the luck of when one retires has the biggest effect of all, but there is no effective way to mess with that ex ante.
User avatar
canadianbacon
Posts: 677
Joined: Sun Nov 10, 2019 9:04 pm

Re: Stock/Bond Mix in Retirement: The Surprise In The Data

Post by canadianbacon »

Northern Flicker wrote: Mon Oct 25, 2021 1:56 am
willthrill81 wrote: Again, if stocks do well when you're heavily tilted to bonds early in retirement, you'll come out behind with a bond tent strategy. And if bonds do poorly during that decade, you're doubly hosed.
While I am not advocating for a bond tent strategy, you are implicitly using a metric of highest return in evaluating the strategy. Sure, there are scenarios where a bond tent has a lower return. The question is, however, how does a bond tent affect the probability of running out of money?

I don't think that question has ever been answered.
This. A bond tent is not about maximizing terminal balance. It's insurance. You should expect to pay for insurance.
Bulls make money, bears make money, pigs get slaughtered.
Northern Flicker
Posts: 15369
Joined: Fri Apr 10, 2015 12:29 am

Re: Stock/Bond Mix in Retirement: The Surprise In The Data

Post by Northern Flicker »

The elephant in the room is inflation-indexed bonds. TIPS and series I savings bonds did not exist for the overwhelming number of calendar years available for SWR simulations. A low equity allocation and high nominal bond allocation lowers sequence of return risk but increases inflation risk. Incorporating inflation-indexed bonds into the allocation is a way of bridging the two goals (managing both risks) but we don't have much historical data to look at.

The historical bias I see in the bond tent samples is that the period that would challenge a low equity allocation with a high allocation to nominal bonds was 1967-1981. Bad retirement start years for this period in Bengen's sample in the original work on this from 1994 were 1965, 1966, 1968, and 1969. If a bond tent were used, there was a bear market in stocks 1973-1974, early in the period, followed by a not particular robust recovery. The bond tent's weakness to cope with inflation was rescued by the downturn being early in the period when the equity allocation was lower, with a weak recovery when the equity allocation was higher. The bond tent strategy led to rebalancing to a higher asset allocation during the bear market.

A question then is can we count on inflationary periods behaving that way? What if equities stay roughly flat in real terms while inflation erodes the fixed income portfolio, with withdrawals shrinking the portfolio over time? Then increasing the equity allocation of the now smaller portfolio risks finally getting hit with a bear market and possible portfolio ruin.

I'm not suggesting this is how things should play out, but the point is that a bond tent may not even lower the probability of running out of money. The strategy also may amplify longevity risk due to the above scenario.
Last edited by Northern Flicker on Mon Oct 25, 2021 5:15 pm, edited 1 time in total.
User avatar
willthrill81
Posts: 32250
Joined: Thu Jan 26, 2017 2:17 pm
Location: USA
Contact:

Re: Stock/Bond Mix in Retirement: The Surprise In The Data

Post by willthrill81 »

canadianbacon wrote: Mon Oct 25, 2021 1:56 pm
Northern Flicker wrote: Mon Oct 25, 2021 1:56 am
willthrill81 wrote: Again, if stocks do well when you're heavily tilted to bonds early in retirement, you'll come out behind with a bond tent strategy. And if bonds do poorly during that decade, you're doubly hosed.
While I am not advocating for a bond tent strategy, you are implicitly using a metric of highest return in evaluating the strategy. Sure, there are scenarios where a bond tent has a lower return. The question is, however, how does a bond tent affect the probability of running out of money?

I don't think that question has ever been answered.
This. A bond tent is not about maximizing terminal balance. It's insurance. You should expect to pay for insurance.
But you want some assurance that the 'insurance' will pay off when you need it.
The Sensible Steward
User avatar
Topic Author
martincmartin
Posts: 900
Joined: Wed Jul 02, 2014 3:04 pm
Location: Boston, MA USA

Re: Stock/Bond Mix in Retirement: The Surprise In The Data

Post by martincmartin »

Northern Flicker wrote: Mon Oct 25, 2021 3:39 pm The elephant in the room is inflation-indexed bonds.
True. In fact, there are many others: international stocks & bonds, REITs, Small Cap Value, the list goes on. The analysis I did was just Large Cap Blend vs Total Bond Market. (My next post: Short Term Treasuries do significantly better than TBM and even ITT.)
1967-1981.
Totally agree, see my latest post.
Bad retirement start years ... were 1965, 1966, 1968, and 1969.
I used all start years from 1871 to 1991. That's 130 different years.
If a bond tent were used, there was a bear market in stocks 1973-1974, early in the period, followed by a not particular robust recovery. The bond tent's weakness to cope with inflation was rescued by the downturn being early in the period when the equity allocation was lower, with a weak recovery when the equity allocation was lower. The bond tent strategy led to rebalancing to a higher asset allocation during the bear market.
The optimization of bond/stock mix works by tweaking things to improve the worst period, 1966 - 1995, until it has improved that period and hurt another period, to the point where the other period is now worse. So in the end, the optimizer makes the worst and second worst periods end with the same final value. For 4% SWR 30 years LCB & TBM, that other period is 1906 - 1935. WWI had inflation much worse than the 1970s (although didn't last nearly as long, and there was a strong recovery after, etc.) So, to your point: it doesn't fit to just 1966, but balances it with another period, 1906.

I think your point is still valid. A large part of what it does is find the period with the worst returns, then tweak itself to fit to exactly that period for a while. So it could very well be overfitting to peculiarities of that period, and there's no guarantee that the worst over the next 30 years looks anything like that.
A question then is can we count on inflationary periods behaving that way? What if equities stay roughly flat in real terms while inflation erodes the fixed income portfolio, with withdrawals shrinking the portfolio over time? Then increasing the equity allocation of the now smaller portfolio risks finally getting hit with a bear market and possible portfolio ruin.
In the end, I'm retiring soon, have a 50/50 mix, and intend to increase it by 2% a year to 70/30, then stay at 70/30 indefinitely. So in your scenario, annual rebalancing helps.
The strategy also may amplify longevity risk due to the above scenario.
I wish they still made CPI linked SPIAs.
secondopinion
Posts: 6011
Joined: Wed Dec 02, 2020 12:18 pm

Re: Stock/Bond Mix in Retirement: The Surprise In The Data

Post by secondopinion »

willthrill81 wrote: Mon Oct 25, 2021 5:06 pm
canadianbacon wrote: Mon Oct 25, 2021 1:56 pm
Northern Flicker wrote: Mon Oct 25, 2021 1:56 am
willthrill81 wrote: Again, if stocks do well when you're heavily tilted to bonds early in retirement, you'll come out behind with a bond tent strategy. And if bonds do poorly during that decade, you're doubly hosed.
While I am not advocating for a bond tent strategy, you are implicitly using a metric of highest return in evaluating the strategy. Sure, there are scenarios where a bond tent has a lower return. The question is, however, how does a bond tent affect the probability of running out of money?

I don't think that question has ever been answered.
This. A bond tent is not about maximizing terminal balance. It's insurance. You should expect to pay for insurance.
But you want some assurance that the 'insurance' will pay off when you need it.
Right. I would never pay for insurance that will not save me when I need it the most.

For example, I cannot take the risk of long-term disability (as I am single and lack a major support network); so, I have long-term disability insurance (and quite a bit of it). In fact, I could become disabled tomorrow and I would likely be able to retire the rest of my entire life (my after-tax pay right now almost equals tax-free insurance benefit until about 65 [I do not think it is inflation adjusted, unfortunately]; smart investing and strong savings rate would likely allow me to make up the rest of my life, however, probably to age 90-95). But course, I would like to eat lobster once in a while; so I hope I never need it.

But yes, hedge out what will actually help you the most to hedge out. The biggest bang for the buck.
Passive investing: not about making big bucks but making profits. Active investing: not about beating the market but meeting goals. Speculation: not about timing the market but taking profitable risks.
secondopinion
Posts: 6011
Joined: Wed Dec 02, 2020 12:18 pm

Re: Stock/Bond Mix in Retirement: The Surprise In The Data

Post by secondopinion »

martincmartin wrote: Mon Oct 25, 2021 5:24 pm
Northern Flicker wrote: Mon Oct 25, 2021 3:39 pm The elephant in the room is inflation-indexed bonds.
True. In fact, there are many others: international stocks & bonds, REITs, Small Cap Value, the list goes on. The analysis I did was just Large Cap Blend vs Total Bond Market. (My next post: Short Term Treasuries do significantly better than TBM and even ITT.)
So cash-similar bonds (but not cash-like) actually have a purpose? I guess I knew it already; but yes, they are unique in comparison to bonds of longer duration as they behave closer to cash instead.
Passive investing: not about making big bucks but making profits. Active investing: not about beating the market but meeting goals. Speculation: not about timing the market but taking profitable risks.
User avatar
Ben Mathew
Posts: 2720
Joined: Tue Mar 13, 2018 11:41 am
Location: Seattle

Re: Stock/Bond Mix in Retirement: The Surprise In The Data

Post by Ben Mathew »

willthrill81 wrote: Mon Oct 25, 2021 10:11 am As such, we must question whether these results were mere historic artifacts that are not likely to be replicated going forward.
Northern Flicker wrote: Mon Oct 25, 2021 3:39 pm The historical bias I see in the bond tent samples is that the period that would challenge a low equity allocation with a high allocation to nominal bonds was 1967-1981. Bad retirement start years for this period in Bengen's sample in the original work on this from 1994 were 1965, 1966, 1968, and 1969. If a bond tent were used, there was a bear market in stocks 1973-1974, early in the period, followed by a not particular robust recovery. The bond tent's weakness to cope with inflation was rescued by the downturn being early in the period when the equity allocation was lower, with a weak recovery when the equity allocation was higher. The bond tent strategy led to rebalancing to a higher asset allocation during the bear market.
Monte Carlo simulations drawing randomly from the historical distribution (with or without replacement) will remove these timing artifatcts. If it hasn't been done already, it would be useful to do a Monte Carlo simulation and see if the upward sloping glidepath result survives.
Total Portfolio Allocation and Withdrawal (TPAW)
User avatar
willthrill81
Posts: 32250
Joined: Thu Jan 26, 2017 2:17 pm
Location: USA
Contact:

Re: Stock/Bond Mix in Retirement: The Surprise In The Data

Post by willthrill81 »

Ben Mathew wrote: Mon Oct 25, 2021 6:12 pm
willthrill81 wrote: Mon Oct 25, 2021 10:11 am As such, we must question whether these results were mere historic artifacts that are not likely to be replicated going forward.
Northern Flicker wrote: Mon Oct 25, 2021 3:39 pm The historical bias I see in the bond tent samples is that the period that would challenge a low equity allocation with a high allocation to nominal bonds was 1967-1981. Bad retirement start years for this period in Bengen's sample in the original work on this from 1994 were 1965, 1966, 1968, and 1969. If a bond tent were used, there was a bear market in stocks 1973-1974, early in the period, followed by a not particular robust recovery. The bond tent's weakness to cope with inflation was rescued by the downturn being early in the period when the equity allocation was lower, with a weak recovery when the equity allocation was higher. The bond tent strategy led to rebalancing to a higher asset allocation during the bear market.
Monte Carlo simulations drawing randomly from the historical distribution (with or without replacement) will remove these timing artifatcts. If it hasn't been done already, it would be useful to do a Monte Carlo simulation and see if the upward sloping glidepath result survives.
Yes, MC simulations would help, although they bring their own issues to the mix, in particular, 'fat' tails. Using 'blocks' of years rather than individual years helps to correct this problem. I've not seen any such analyses with bond tent strategies.
The Sensible Steward
User avatar
Ben Mathew
Posts: 2720
Joined: Tue Mar 13, 2018 11:41 am
Location: Seattle

Re: Stock/Bond Mix in Retirement: The Surprise In The Data

Post by Ben Mathew »

willthrill81 wrote: Mon Oct 25, 2021 6:26 pm
Ben Mathew wrote: Mon Oct 25, 2021 6:12 pm
willthrill81 wrote: Mon Oct 25, 2021 10:11 am As such, we must question whether these results were mere historic artifacts that are not likely to be replicated going forward.
Northern Flicker wrote: Mon Oct 25, 2021 3:39 pm The historical bias I see in the bond tent samples is that the period that would challenge a low equity allocation with a high allocation to nominal bonds was 1967-1981. Bad retirement start years for this period in Bengen's sample in the original work on this from 1994 were 1965, 1966, 1968, and 1969. If a bond tent were used, there was a bear market in stocks 1973-1974, early in the period, followed by a not particular robust recovery. The bond tent's weakness to cope with inflation was rescued by the downturn being early in the period when the equity allocation was lower, with a weak recovery when the equity allocation was higher. The bond tent strategy led to rebalancing to a higher asset allocation during the bear market.
Monte Carlo simulations drawing randomly from the historical distribution (with or without replacement) will remove these timing artifatcts. If it hasn't been done already, it would be useful to do a Monte Carlo simulation and see if the upward sloping glidepath result survives.
Yes, MC simulations would help, although they bring their own issues to the mix, in particular, 'fat' tails. Using 'blocks' of years rather than individual years helps to correct this problem. I've not seen any such analyses with bond tent strategies.
Why would fat tails create a problem for Monte Carlo simulations?
Total Portfolio Allocation and Withdrawal (TPAW)
User avatar
willthrill81
Posts: 32250
Joined: Thu Jan 26, 2017 2:17 pm
Location: USA
Contact:

Re: Stock/Bond Mix in Retirement: The Surprise In The Data

Post by willthrill81 »

Ben Mathew wrote: Mon Oct 25, 2021 6:51 pm
willthrill81 wrote: Mon Oct 25, 2021 6:26 pm
Ben Mathew wrote: Mon Oct 25, 2021 6:12 pm
willthrill81 wrote: Mon Oct 25, 2021 10:11 am As such, we must question whether these results were mere historic artifacts that are not likely to be replicated going forward.
Northern Flicker wrote: Mon Oct 25, 2021 3:39 pm The historical bias I see in the bond tent samples is that the period that would challenge a low equity allocation with a high allocation to nominal bonds was 1967-1981. Bad retirement start years for this period in Bengen's sample in the original work on this from 1994 were 1965, 1966, 1968, and 1969. If a bond tent were used, there was a bear market in stocks 1973-1974, early in the period, followed by a not particular robust recovery. The bond tent's weakness to cope with inflation was rescued by the downturn being early in the period when the equity allocation was lower, with a weak recovery when the equity allocation was higher. The bond tent strategy led to rebalancing to a higher asset allocation during the bear market.
Monte Carlo simulations drawing randomly from the historical distribution (with or without replacement) will remove these timing artifatcts. If it hasn't been done already, it would be useful to do a Monte Carlo simulation and see if the upward sloping glidepath result survives.
Yes, MC simulations would help, although they bring their own issues to the mix, in particular, 'fat' tails. Using 'blocks' of years rather than individual years helps to correct this problem. I've not seen any such analyses with bond tent strategies.
Why would fat tails create a problem for Monte Carlo simulations?
Because the fat tails are probably very unrealistic. Not many plans could tolerate something like the Great Depression followed by 1970s stagflation followed by the Great Recession, etc., but that's precisely what some MC simulations will randomly create. Derek Tharp wrote a nice piece on Kitces' website about this issue.

There are ways to at least partly combat this problem, but they aren't always implemented.
The Sensible Steward
User avatar
canadianbacon
Posts: 677
Joined: Sun Nov 10, 2019 9:04 pm

Re: Stock/Bond Mix in Retirement: The Surprise In The Data

Post by canadianbacon »

willthrill81 wrote: Mon Oct 25, 2021 5:06 pm
canadianbacon wrote: Mon Oct 25, 2021 1:56 pm
Northern Flicker wrote: Mon Oct 25, 2021 1:56 am
willthrill81 wrote: Again, if stocks do well when you're heavily tilted to bonds early in retirement, you'll come out behind with a bond tent strategy. And if bonds do poorly during that decade, you're doubly hosed.
While I am not advocating for a bond tent strategy, you are implicitly using a metric of highest return in evaluating the strategy. Sure, there are scenarios where a bond tent has a lower return. The question is, however, how does a bond tent affect the probability of running out of money?

I don't think that question has ever been answered.
This. A bond tent is not about maximizing terminal balance. It's insurance. You should expect to pay for insurance.
But you want some assurance that the 'insurance' will pay off when you need it.
And a rising equity glidepath does, in the event of an adverse sequence. You already linked to ERN's work, although you seemed to minimize or dismiss the results as noise.
Bulls make money, bears make money, pigs get slaughtered.
User avatar
willthrill81
Posts: 32250
Joined: Thu Jan 26, 2017 2:17 pm
Location: USA
Contact:

Re: Stock/Bond Mix in Retirement: The Surprise In The Data

Post by willthrill81 »

canadianbacon wrote: Mon Oct 25, 2021 7:33 pm
willthrill81 wrote: Mon Oct 25, 2021 5:06 pm
canadianbacon wrote: Mon Oct 25, 2021 1:56 pm
Northern Flicker wrote: Mon Oct 25, 2021 1:56 am
willthrill81 wrote: Again, if stocks do well when you're heavily tilted to bonds early in retirement, you'll come out behind with a bond tent strategy. And if bonds do poorly during that decade, you're doubly hosed.
While I am not advocating for a bond tent strategy, you are implicitly using a metric of highest return in evaluating the strategy. Sure, there are scenarios where a bond tent has a lower return. The question is, however, how does a bond tent affect the probability of running out of money?

I don't think that question has ever been answered.
This. A bond tent is not about maximizing terminal balance. It's insurance. You should expect to pay for insurance.
But you want some assurance that the 'insurance' will pay off when you need it.
And a rising equity glidepath does, in the event of an adverse sequence. You already linked to ERN's work, although you seemed to minimize or dismiss the results as noise.
The benefits were not consistent and, on average, were very small. If you want 'insurance' against a bad sequence of returns, there are several approaches that have provided much larger and more consistent benefits than a rising equity glidepath.
The Sensible Steward
Post Reply