A bond rule of thumb better than "age in bonds"

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
User avatar
Topic Author
vineviz
Posts: 9933
Joined: Tue May 15, 2018 1:55 pm

A bond rule of thumb better than "age in bonds"

Post by vineviz »

Asset allocation experts generally agree that the classic "age in bonds" rule of thumb is pretty terrible investment guidance. It produces suggested asset allocations that bear virtually no resemblance to allocations constructed using more empirically and/or theoretically sound methods.

The main advantage it has, IMHO, is that it easy to remember and easy to calculate. Simple modifications like "age minus 20" produce much better results albeit ones that are still recklessly conservative for midde-aged retirement investors (e.g. ages 35 to 50).

I've been trying to come up with an easily computable rule that investors can use to set their own glide path, and I think this will do it.

Bonds = 2 * (age - 40)

Most investors will eschew leverage, so this basically amounts to 100/0 allocations until age 40. This is consistent with the glide paths of most target date funds. In the following graph you can see a comparison of the (terrible) "age in bonds" rule, the (less terrible) "age minus 20" rule, my (even less terrible) "2x(age-40)" rule, and the computed "optimal" age-based glide path.

Image
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch
rbaldini
Posts: 1436
Joined: Mon Mar 23, 2015 3:20 pm

Re: A bond rule of thumb better than "age in bonds"

Post by rbaldini »

Briefly, how is the "optimal" bond allocation determined? Surely this is some average - not the same for every investor.
User avatar
anon_investor
Posts: 6185
Joined: Mon Jun 03, 2019 1:43 pm

Re: A bond rule of thumb better than "age in bonds"

Post by anon_investor »

So 100% equities if 40 or under? That actually makes a lot of sense. So 100% LLT for fixed income until over the age of 50?
Glockenspiel
Posts: 1164
Joined: Thu Feb 08, 2018 1:20 pm

Re: A bond rule of thumb better than "age in bonds"

Post by Glockenspiel »

I really like this. I'm 35 and have actually been using Age minus 20 in bonds, but it does feel conservative to have 20% bonds at age 40, when I'll still be 20 years away from retirement. I do wonder how the "optimal" bond allocation in the graph was computed, though.
HawkeyePierce
Posts: 1828
Joined: Tue Mar 05, 2019 10:29 pm
Location: Colorado

Re: A bond rule of thumb better than "age in bonds"

Post by HawkeyePierce »

I agree this may be mathematically optimal but I'm not convinced it's behaviorally optimal.

(Behaviorally, I still think the vast vast majority of investors should be in a target-date or target-allocation fund, but anyways)

I'm 30 and I hold 80/20. I think Bernstein was generally correct with his guidance of "never more than 75% or less than 25% bonds".
User avatar
Topic Author
vineviz
Posts: 9933
Joined: Tue May 15, 2018 1:55 pm

Re: A bond rule of thumb better than "age in bonds"

Post by vineviz »

For those who prefer to think in terms of wealth, rather than age, there's a roughly equivalent wealth-based rule of thumb.

Compute your expected retirement withdrawals (in current dollars) as a multiple of your current retirement savings: this is (in percentage terms) your bond allocation.

So, for example, imagine that your current retirement savaging are $200,000 and you expect your annual portfolio withdrawals in retirement will be $20,000. 200,000 divided by 20,000 is 10. Your bond allocation is 10%.

If your current retirement savaging are $800,000 and you expect your annual portfolio withdrawals in retirement will be $20,000 then your bond allocation is 40% ( 800,000 divided by 20,000 is 40).

This will produce a slightly different glide path than the age-based rule, but the shape of the glide path will be roughly similar for most investors.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch
User avatar
David Jay
Posts: 10765
Joined: Mon Mar 30, 2015 5:54 am
Location: Michigan

Re: A bond rule of thumb better than "age in bonds"

Post by David Jay »

HawkeyePierce wrote: Thu May 07, 2020 9:53 am...I'm not convinced it's behaviorally optimal.
After watching this board over the last 3 months, I will go a step further: This is definitely not behaviorally optimal.

We have re-discovered that AA is intensely personal. Bernstein suggests a 40% range of based on risk tolerance ("Investor's Manifesto"), i.e. at an age where the calculation suggests 70/30, a highly risk tolerant individual can be 90/10 but a risk averse individual may need to be 50/50.
Last edited by David Jay on Thu May 07, 2020 10:09 am, edited 2 times in total.
Prediction is very difficult, especially about the future - Niels Bohr | To get the "risk premium", you really do have to take the risk - nisiprius
rockstar
Posts: 1624
Joined: Mon Feb 03, 2020 6:51 pm

Re: A bond rule of thumb better than "age in bonds"

Post by rockstar »

But what type of bonds and at what duration? It’s hard to find yield that will keep up with inflation right now.
csmath
Posts: 563
Joined: Sat Oct 13, 2018 11:32 am

Re: A bond rule of thumb better than "age in bonds"

Post by csmath »

As you acknowledge, it isn't optimal. However, I think it is a huge improvement over "age in bonds" as a baseline from which to start personalizing from.
rbaldini
Posts: 1436
Joined: Mon Mar 23, 2015 3:20 pm

Re: A bond rule of thumb better than "age in bonds"

Post by rbaldini »

csmath wrote: Thu May 07, 2020 10:09 am As you acknowledge, it isn't optimal. However, I think it is a huge improvement over "age in bonds" as a baseline from which to start personalizing from.
Again I am left wondering how the optimal strategy was determined.

EDIT: never mind, addressed below
Last edited by rbaldini on Thu May 07, 2020 10:11 am, edited 1 time in total.
User avatar
Topic Author
vineviz
Posts: 9933
Joined: Tue May 15, 2018 1:55 pm

Re: A bond rule of thumb better than "age in bonds"

Post by vineviz »

rbaldini wrote: Thu May 07, 2020 9:42 am Briefly, how is the "optimal" bond allocation determined? Surely this is some average - not the same for every investor.
What I call "optimal" here doesn't directly take investor risk tolerance into account, so that's a wrinkle for sure. Instead, it is a mathematically optimal glide path that is intended to minimize the dispersion of terminal wealth at the date of retirement.

In short, the starting assumptions are an investor who starts saving for retirement at age 25 by putting 15% of their income into a retirement account. Each year, their wages (and therefore their retirement contribution) increases to match the rate of inflation.

The goal, then, becomes achieving a portfolio at retirement that has an expected value which is some multiple of the retirement withdrawal. E.g. 25x spending or 30x spending.

The optimal glide path is the one that produces that expected value with the smallest possible variance. In other words, my goal is to have 30x retirement spending, and I want a glide path that gets me close to that as possible as often as possible (probabilistically speaking).

This effectively treats the process as having two discrete stages: accumulation is stage one, and decumulation is stage two.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch
csmath
Posts: 563
Joined: Sat Oct 13, 2018 11:32 am

Re: A bond rule of thumb better than "age in bonds"

Post by csmath »

rbaldini wrote: Thu May 07, 2020 10:10 am
csmath wrote: Thu May 07, 2020 10:09 am As you acknowledge, it isn't optimal. However, I think it is a huge improvement over "age in bonds" as a baseline from which to start personalizing from.
Again I am left wondering how the optimal strategy was determined.

EDIT: never mind, addressed below
Yeah, usually you really can't have both "optimal" AND "rule of thumb". That is what the "start personalizing from" part was about; trying to get from the "rule of thumb" part to the "optimal" part since that is likely different for everyone.
rbaldini
Posts: 1436
Joined: Mon Mar 23, 2015 3:20 pm

Re: A bond rule of thumb better than "age in bonds"

Post by rbaldini »

vineviz wrote: Thu May 07, 2020 10:10 am
rbaldini wrote: Thu May 07, 2020 9:42 am Briefly, how is the "optimal" bond allocation determined? Surely this is some average - not the same for every investor.
What I call "optimal" here doesn't directly take investor risk tolerance into account, so that's a wrinkle for sure. Instead, it is a mathematically optimal glide path that is intended to minimize the dispersion of terminal wealth at the date of retirement.

In short, the starting assumptions are an investor who starts saving for retirement at age 25 by putting 15% of their income into a retirement account. Each year, their wages (and therefore their retirement contribution) increases to match the rate of inflation.

The goal, then, becomes achieving a portfolio at retirement that has an expected value which is some multiple of the retirement withdrawal. E.g. 25x spending or 30x spending.

The optimal glide path is the one that produces that expected value with the smallest possible variance. In other words, my goal is to have 30x retirement spending, and I want a glide path that gets me close to that as possible as often as possible.

This effectively treats the process as having two discrete stages: accumulation is stage one, and decumulation is stage two.
Thanks for sharing.

Here's an approach that I think would provide an interesting alternative: choose a strategy that maximizes expected return, subject to the constraint of providing a chance of at least x% of achieving a retirement net worth of $y. In other words, the constraint here is a quantile of the final distribution, not the expected value. The motivation is "I want to be reasonably certain (x%) of having at least enough money to retire on ($y), but of course more money is better (maximize return)". Obviously the final answer would depend on the values selected, the starting value, etc.

One implication of this approach would be that, all else equal, if you have more money now, you have afford to lose more, and therefore you can afford to take more risk and invest more in stock. I'm not sure if it would produce a simple age-based rule. In fact, I'm fairly certain that, whether it's your model or mine, age is not the only variable. But then, these are just rules of thumb.
Last edited by rbaldini on Thu May 07, 2020 10:21 am, edited 1 time in total.
rkhusky
Posts: 10909
Joined: Thu Aug 18, 2011 8:09 pm

Re: A bond rule of thumb better than "age in bonds"

Post by rkhusky »

Optimality is only known in hindsight.

Age in bonds is fine for a conservative investor. (In my opinion. And any other judgement of its suitability is also just an opinion.)

Age-20 is fine for a moderately aggressive investor.

Time until retirement is a better measure than age.
terran
Posts: 2038
Joined: Sat Jan 10, 2015 10:50 pm

Re: A bond rule of thumb better than "age in bonds"

Post by terran »

Interesting. I'd be interested how this could be adapted to early retirees: steeper/higher savings rates, and shorter accumulation periods. Is there anything more you can share that would help one figure out the optimal path under such altered conditions?
TheoLeo
Posts: 322
Joined: Mon Apr 02, 2018 11:39 am

Re: A bond rule of thumb better than "age in bonds"

Post by TheoLeo »

David Jay wrote: Thu May 07, 2020 9:59 am
HawkeyePierce wrote: Thu May 07, 2020 9:53 am...I'm not convinced it's behaviorally optimal.
After watching this board over the last 3 months, I will go a step further: This is definitely not behaviorally optimal.

We have re-discovered that AA is intensely personal. Bernstein suggests a 40% range of based on risk tolerance ("Investor's Manifesto"), i.e. at an age where the calculation suggests 70/30, a highly risk tolerant individual can be 90/10 but a risk averse individual may need to be 50/50.
I think behavior is not only affected by the magnitude of the the drop in value of ones portfolio during a crash. What might be even more important is if one has faith in the asset allocation to dig itself out of this hole again. If your supposedly safe bond allocation goes down, and further down, and further down, maybe in a rising rate environment, then this creates a lot of uncertainty that people might not have with stocks. It is easier to put faith in the ability of companies to create profits over time than to put faith in central banks to stick to a certain policy.
User avatar
midareff
Posts: 7536
Joined: Mon Nov 29, 2010 10:43 am
Location: Biscayne Bay, South Florida

Re: A bond rule of thumb better than "age in bonds"

Post by midareff »

Bonds = 2 * (age - 40) Let me see..... age 72, minus 40 = 32*2 = 64% bonds... it doesn't work for me. I'm at 52% bonds and 4% cash. The ten year treasury is yielding .721% and the rolling 12 month CPI-U is at 2.00%. Your plan would have me at 36% equities which most Gurus expect to yield 5%-6% over the next decade with the Shiller still at 27.05.... so what we have here is 36% of the portfolio at maybe 3.5% real and 64% at -1.3% real, unless inflation rises and bonds get crushed.

It may be better for certain age groups, it doesn't work for me.
User avatar
Ben Mathew
Posts: 1570
Joined: Tue Mar 13, 2018 11:41 am
Location: Seattle

Re: A bond rule of thumb better than "age in bonds"

Post by Ben Mathew »

vineviz wrote: Thu May 07, 2020 10:10 am What I call "optimal" here doesn't directly take investor risk tolerance into account, so that's a wrinkle for sure. Instead, it is a mathematically optimal glide path that is intended to minimize the dispersion of terminal wealth at the date of retirement.

In short, the starting assumptions are an investor who starts saving for retirement at age 25 by putting 15% of their income into a retirement account. Each year, their wages (and therefore their retirement contribution) increases to match the rate of inflation.

The goal, then, becomes achieving a portfolio at retirement that has an expected value which is some multiple of the retirement withdrawal. E.g. 25x spending or 30x spending.

The optimal glide path is the one that produces that expected value with the smallest possible variance. In other words, my goal is to have 30x retirement spending, and I want a glide path that gets me close to that as possible as often as possible (probabilistically speaking).

This effectively treats the process as having two discrete stages: accumulation is stage one, and decumulation is stage two.
Should be a better fit than age in bonds for most people I think.

The solution to the utility maximization problem with CRRA utility is constant fraction of lifetime wealth in stocks. That implies a downward glide path till retirement and flat after retirement. Adding a no-leverage condition cuts off the leverage when young and make it flat at 100% stocks when young. Your proposed glide path seems to have the same basic shape till retirement.

Might be useful to tie the age when 100% stocks end to risk preference, savings rates and age of retirement. More complex, but more tailored.
User avatar
KidIcarus
Posts: 18
Joined: Sun Feb 04, 2018 12:40 pm

Re: A bond rule of thumb better than "age in bonds"

Post by KidIcarus »

vineviz wrote: Thu May 07, 2020 9:58 am For those who prefer to think in terms of wealth, rather than age, there's a roughly equivalent wealth-based rule of thumb.

Compute your expected retirement withdrawals (in current dollars) as a multiple of your current retirement savings: this is (in percentage terms) your bond allocation.

So, for example, imagine that your current retirement savaging are $200,000 and you expect your annual portfolio withdrawals in retirement will be $20,000. 200,000 divided by 20,000 is 10. Your bond allocation is 10%.

If your current retirement savaging are $800,000 and you expect your annual portfolio withdrawals in retirement will be $20,000 then your bond allocation is 40% ( 800,000 divided by 20,000 is 40).

This will produce a slightly different glide path than the age-based rule, but the shape of the glide path will be roughly similar for most investors.
Can you explain the rational behind this? As you accumulate more wealth relative to expenses, your bond percentage will go down?
User avatar
Munir
Posts: 2813
Joined: Mon Feb 26, 2007 4:39 pm
Location: Oregon

Re: A bond rule of thumb better than "age in bonds"

Post by Munir »

The graphs stop at age 65. Why?
User avatar
Topic Author
vineviz
Posts: 9933
Joined: Tue May 15, 2018 1:55 pm

Re: A bond rule of thumb better than "age in bonds"

Post by vineviz »

midareff wrote: Thu May 07, 2020 10:39 am Bonds = 2 * (age - 40) Let me see..... age 72, minus 40 = 32*2 = 64% bonds... it doesn't work for me. I'm at 52% bonds and 4% cash. The ten year treasury is yielding .721% and the rolling 12 month CPI-U is at 2.00%. Your plan would have me at 36% equities which most Gurus expect to yield 5%-6% over the next decade with the Shiller still at 27.05.... so what we have here is 36% of the portfolio at maybe 3.5% real and 64% at -1.3% real, unless inflation rises and bonds get crushed.

It may be better for certain age groups, it doesn't work for me.

You can always use “age in bonds”, I guess.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch
User avatar
vitaflo
Posts: 1572
Joined: Sat Sep 03, 2011 3:02 pm

Re: A bond rule of thumb better than "age in bonds"

Post by vitaflo »

An even better rule of thumb is not to use age at all. Age is just a proxy for portfolio size, specifically how much you have in your portfolio vs what you need in your portfolio to retire on.

So just use that instead. The math is more complicated, but it means your glide path will reach your desired AA regardless of when you retire, which is especially useful for those who want to retire early.
grok87
Posts: 9476
Joined: Tue Feb 27, 2007 9:00 pm

Re: A bond rule of thumb better than "age in bonds"

Post by grok87 »

this thread may be relevant
viewtopic.php?f=10&t=311560
it suggests, roughly, at retirement (say age 65 for example) to be say 42% in bonds/cash.

10 years before retirement one would be 30% in bonds.
RIP Mr. Bogle.
User avatar
Topic Author
vineviz
Posts: 9933
Joined: Tue May 15, 2018 1:55 pm

Re: A bond rule of thumb better than "age in bonds"

Post by vineviz »

Munir wrote: Thu May 07, 2020 10:52 am The graphs stop at age 65. Why?
terran wrote: Thu May 07, 2020 10:29 am Interesting. I'd be interested how this could be adapted to early retirees: steeper/higher savings rates, and shorter accumulation periods. Is there anything more you can share that would help one figure out the optimal path under such altered conditions?
I hinted at this earlier, but saving for retirement (accumulation) and spending in retirement (decumulation) are very different problems.

Accumulation is essentially an investment "problem" but decumulation isn't: it's an income "problem", which requires a distinct approach that was beyond the scope of the OP.

This new rule of thumb is designed for people who are turning income into wealth. People who are turning wealth into income (i.e. retirees) probably shouldn't be relying on asset allocation as their primary method of control.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch
AlohaBill
Posts: 150
Joined: Thu Jan 24, 2008 1:20 pm
Location: California

Re: A bond rule of thumb better than "age in bonds"

Post by AlohaBill »

Who is this rule for? What type of investor should heed your call? C’mon are you just a little bit trolling the bogleheads here? Lol. Is this really for everyone? 100% stock for 40 years!? How many investors can or would want to tak that kind of risk?
User avatar
midareff
Posts: 7536
Joined: Mon Nov 29, 2010 10:43 am
Location: Biscayne Bay, South Florida

Re: A bond rule of thumb better than "age in bonds"

Post by midareff »

vineviz wrote: Thu May 07, 2020 10:58 am
midareff wrote: Thu May 07, 2020 10:39 am Bonds = 2 * (age - 40) Let me see..... age 72, minus 40 = 32*2 = 64% bonds... it doesn't work for me. I'm at 52% bonds and 4% cash. The ten year treasury is yielding .721% and the rolling 12 month CPI-U is at 2.00%. Your plan would have me at 36% equities which most Gurus expect to yield 5%-6% over the next decade with the Shiller still at 27.05.... so what we have here is 36% of the portfolio at maybe 3.5% real and 64% at -1.3% real, unless inflation rises and bonds get crushed.

It may be better for certain age groups, it doesn't work for me.

You can always use “age in bonds”, I guess.
I've been maintaining a relatively balanced allocation of 48/48/4 .. equities, bonds and cash for the last 17 years or so. The run of Treasuries and such make them difficult for a retiree to own. With a current .721% yield on the ten year and inflation a a rolling 12 month of 2% CDs and Ally Savings at 1.5% look better.. so do corporates and munis ranging from 2.4% to 3.2%. If you die soon none of it matters much, if you live a long life some real return is required from that portion of your portfolio.
User avatar
willthrill81
Posts: 24608
Joined: Thu Jan 26, 2017 3:17 pm
Location: USA

Re: A bond rule of thumb better than "age in bonds"

Post by willthrill81 »

midareff wrote: Thu May 07, 2020 10:39 am Bonds = 2 * (age - 40) Let me see..... age 72, minus 40 = 32*2 = 64% bonds... it doesn't work for me.
Vineviz's formula is probably best suited for accumulators, not those in the withdrawal phase. Presumably, you would only use the formula to increase your bond allocation until it reached the AA you wanted to at least begin your retirement at.
“Good and ill have not changed since yesteryear; nor are they one thing among Elves and Dwarves and another among Men.” J.R.R. Tolkien, The Lord of the Rings
csmath
Posts: 563
Joined: Sat Oct 13, 2018 11:32 am

Re: A bond rule of thumb better than "age in bonds"

Post by csmath »

midareff wrote: Thu May 07, 2020 10:39 am Bonds = 2 * (age - 40) Let me see..... age 72, minus 40 = 32*2 = 64% bonds... it doesn't work for me. I'm at 52% bonds and 4% cash. The ten year treasury is yielding .721% and the rolling 12 month CPI-U is at 2.00%. Your plan would have me at 36% equities which most Gurus expect to yield 5%-6% over the next decade with the Shiller still at 27.05.... so what we have here is 36% of the portfolio at maybe 3.5% real and 64% at -1.3% real, unless inflation rises and bonds get crushed.

It may be better for certain age groups, it doesn't work for me.
Or.... you could say the "rule of thumb" part worked great for you... and then you personalized it.

Anytime anyone mentions a "rule of thumb" there is going to be wild disagreement with a ton of specific case application. Maybe we just shouldn't have rules of thumb. Of course generally they are meant for people that really don't know any better let alone where to be begin. Then more knowledgeable people jump in and make the whole thing so complicated that the audience the rules were intended for are completely discouraged and lost.

I'm guilty of doing that myself!
fatcoffeedrinker
Posts: 299
Joined: Mon Mar 23, 2020 2:03 pm

Re: A bond rule of thumb better than "age in bonds"

Post by fatcoffeedrinker »

willthrill81 wrote: Thu May 07, 2020 12:07 pm
midareff wrote: Thu May 07, 2020 10:39 am Bonds = 2 * (age - 40) Let me see..... age 72, minus 40 = 32*2 = 64% bonds... it doesn't work for me.
Vineviz's formula is probably best suited for accumulators, not those in the withdrawal phase. Presumably, you would only use the formula to increase your bond allocation until it reached the AA you wanted to at least begin your retirement at.
Even for accumlators, I think at least two other factors come into play other than age: years until target retirement date and need to take risk. I'm 51 and plan to retire in 3 or so years. I have been very fortunate lately in that my comp has increased to the point where I can now save 5x expenses per year. So for me, I don't want a market crash to push out my retirement date, and because of my savings rate, I don't need much return to hit my retirement goal. That is why I am 40/60 right now. Super conservative for 51, yes, but it works for my situation at this time given my goals. After I retire, I will likely ramp the stock allocation up over time, as I don't want to be 40/60 forever. Maybe end at 60/40 or 65/35, or maybe just keep 10x expenses in FI and invest the rest in stocks -- not quite sure yet.
User avatar
SimpleGift
Posts: 3869
Joined: Tue Feb 08, 2011 3:45 pm
Location: Central Oregon

Re: A bond rule of thumb better than "age in bonds"

Post by SimpleGift »

csmath wrote: Thu May 07, 2020 12:15 pm Anytime anyone mentions a "rule of thumb" there is going to be wild disagreement with a ton of specific case application. Maybe we just shouldn't have rules of thumb.
Totally agree. And I think most Bogleheads would as well, if a poll taken in 2016 of the Bogleheads' actual stock/bond allocations vs. their age is any guide (chart below, n = 217):
From the chart, there are 30-year Bogleheads with a 50% stock allocation and those with 100% stocks. There are 70-year old Bogleheads with 20%-30% stock allocations and those with 80%-100% stocks. In the asset allocation decision, so much depends upon one's individual circumstances, risk tolerances, return goals, etc.
fatcoffeedrinker
Posts: 299
Joined: Mon Mar 23, 2020 2:03 pm

Re: A bond rule of thumb better than "age in bonds"

Post by fatcoffeedrinker »

SimpleGift wrote: Thu May 07, 2020 12:32 pm
csmath wrote: Thu May 07, 2020 12:15 pm Anytime anyone mentions a "rule of thumb" there is going to be wild disagreement with a ton of specific case application. Maybe we just shouldn't have rules of thumb.
Totally agree. And I think most Bogleheads would as well, if a poll taken in 2016 of the Bogleheads' actual stock/bond allocations vs. their age is any guide (chart below, n = 217):
From the chart, there are 30-year Bogleheads with a 50% stock allocation and those with 100% stocks. There are 70-year old Bogleheads with 20%-30% stock allocations and those with 80%-100% stocks. In the asset allocation decision, so much depends upon one's individual circumstances, risk tolerances, return goals, etc.
Interesting. Rough formula based on that graph is bonds = (5*age - 40)/6.

EDIT: Or could be written as (age - 8) * (5/6). So it uses something close to age - 10, but then scales it down to be a little less conservative.
Last edited by fatcoffeedrinker on Thu May 07, 2020 12:49 pm, edited 1 time in total.
User avatar
patrick013
Posts: 3163
Joined: Mon Jul 13, 2015 7:49 pm

Re: A bond rule of thumb better than "age in bonds"

Post by patrick013 »

Don't forget there are basically 3 kinds of investors.
Conservative, moderate, and aggressive. Growth periods
make the aggressive investor look better with no future
regard for decline, deflation, even earthquake. So we
cannot spend variance if and when these events occur but
we can spend an increasing bond AA with only bond risk to
be concerned about.

Inflation wise if incomes exceed expenditures then the
investor is not losing ground budget wise but has excess
income for future inflation concerns. Smaller accounts
might tend to be more aggressive to increase in size to
accomodate a higher desired standard of living even keeping
a permanent low bond AA into retirement.

Middle class and larger account investors make more sense
out of age-in-bonds or perhaps age-in-bonds stop at 50
where the excess budget is favorable, and the excess risk
of a high stock AA does not outweigh the benefits or actual
need when an extended decline of 5 or 10 years is not impossible.

Age-in-bonds acknowledges that growth cannot or may not be
infinite. Others look like they think 8-10% returns are
continuous and mandatory. Without trillion dollar bailouts
(QE) that certainly isn't the case.
age in bonds, buy-and-hold, 10 year business cycle
User avatar
Topic Author
vineviz
Posts: 9933
Joined: Tue May 15, 2018 1:55 pm

Re: A bond rule of thumb better than "age in bonds"

Post by vineviz »

fatcoffeedrinker wrote: Thu May 07, 2020 12:31 pm Even for accumlators, I think at least two other factors come into play other than age: years until target retirement date and need to take risk. I'm 51 and plan to retire in 3 or so years. I have been very fortunate lately in that my comp has increased to the point where I can now save 5x expenses per year. So for me, I don't want a market crash to push out my retirement date, and because of my savings rate, I don't need much return to hit my retirement goal. That is why I am 40/60 right now.
For sure my proposal wasn't intended as a suggestion that a rule of thumb should dominate the decision making process.

In my hierarchy, a really good plan is FAR better than a decent rule of thumb. A decent rule of thumb is FAR better than a terrible rule of thumb. And a terrible rule of thumb is probably better than a completely uninformed guess. ;)

Anyone within 5 years or so of retirement should probably have a really good plan in place for their retirement income, and I presume that you do. That should definitely trump any sort of generalized guidance, which is something it seems like you already know.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch
User avatar
goodenyou
Posts: 2674
Joined: Sun Jan 31, 2010 11:57 pm
Location: Skating to Where the Puck is Going to Be..or on the golf course

Re: A bond rule of thumb better than "age in bonds"

Post by goodenyou »

midareff wrote: Thu May 07, 2020 11:51 am
vineviz wrote: Thu May 07, 2020 10:58 am
midareff wrote: Thu May 07, 2020 10:39 am Bonds = 2 * (age - 40) Let me see..... age 72, minus 40 = 32*2 = 64% bonds... it doesn't work for me. I'm at 52% bonds and 4% cash. The ten year treasury is yielding .721% and the rolling 12 month CPI-U is at 2.00%. Your plan would have me at 36% equities which most Gurus expect to yield 5%-6% over the next decade with the Shiller still at 27.05.... so what we have here is 36% of the portfolio at maybe 3.5% real and 64% at -1.3% real, unless inflation rises and bonds get crushed.

It may be better for certain age groups, it doesn't work for me.

You can always use “age in bonds”, I guess.
I've been maintaining a relatively balanced allocation of 48/48/4 .. equities, bonds and cash for the last 17 years or so. The run of Treasuries and such make them difficult for a retiree to own. With a current .721% yield on the ten year and inflation a a rolling 12 month of 2% CDs and Ally Savings at 1.5% look better.. so do corporates and munis ranging from 2.4% to 3.2%. If you die soon none of it matters much, if you live a long life some real return is required from that portion of your portfolio.
You are making the case for an income floor. This can be achieved with a SPIA where the longevity risk is mitigated.
"Ignorance more frequently begets confidence than does knowledge" | “Do you know how to make a rain dance work? Dance until it rains”
User avatar
HomerJ
Posts: 16468
Joined: Fri Jun 06, 2008 12:50 pm

Re: A bond rule of thumb better than "age in bonds"

Post by HomerJ »

vineviz wrote: Thu May 07, 2020 10:10 amWhat I call "optimal" here doesn't directly take investor risk tolerance into account, so that's a wrinkle for sure.
It's recklessly incomplete then. And that's a serious statement.

At least as serious as yours...
recklessly conservative for midde-aged retirement investors (e.g. ages 35 to 50).
What the heck is "recklessly conservative"?
Last edited by HomerJ on Thu May 07, 2020 12:58 pm, edited 1 time in total.
A Goldman Sachs associate provided a variety of detailed explanations, but then offered a caveat, “If I’m being dead-### honest, though, nobody knows what’s really going on.”
fatcoffeedrinker
Posts: 299
Joined: Mon Mar 23, 2020 2:03 pm

Re: A bond rule of thumb better than "age in bonds"

Post by fatcoffeedrinker »

vineviz wrote: Thu May 07, 2020 12:52 pm
fatcoffeedrinker wrote: Thu May 07, 2020 12:31 pm Even for accumlators, I think at least two other factors come into play other than age: years until target retirement date and need to take risk. I'm 51 and plan to retire in 3 or so years. I have been very fortunate lately in that my comp has increased to the point where I can now save 5x expenses per year. So for me, I don't want a market crash to push out my retirement date, and because of my savings rate, I don't need much return to hit my retirement goal. That is why I am 40/60 right now.
For sure my proposal wasn't intended as a suggestion that a rule of thumb should dominate the decision making process.

In my hierarchy, a really good plan is FAR better than a decent rule of thumb. A decent rule of thumb is FAR better than a terrible rule of thumb. And a terrible rule of thumb is probably better than a completely uninformed guess. ;)

Anyone within 5 years or so of retirement should probably have a really good plan in place for their retirement income, and I presume that you do. That should definitely trump any sort of generalized guidance, which is something it seems like you already know.
Yes, agreed, not a bad rule of thumb for earlier accumulators who still have 5+ years to go.
jrbdmb
Posts: 678
Joined: Tue Oct 06, 2015 4:27 pm

Re: A bond rule of thumb better than "age in bonds"

Post by jrbdmb »

HawkeyePierce wrote: Thu May 07, 2020 9:53 am I agree this may be mathematically optimal but I'm not convinced it's behaviorally optimal.

(Behaviorally, I still think the vast vast majority of investors should be in a target-date or target-allocation fund, but anyways)

I'm 30 and I hold 80/20. I think Bernstein was generally correct with his guidance of "never more than 75% or less than 25% bonds".
I agree that a rule of thumb that recommends 100% stocks is going to be difficult for many investors to follow - in times of market stress this could lead to sub-optimal behavior. Glad that my modified (Age - 20) is only "less terrible." :happy
terran
Posts: 2038
Joined: Sat Jan 10, 2015 10:50 pm

Re: A bond rule of thumb better than "age in bonds"

Post by terran »

vineviz wrote: Thu May 07, 2020 11:07 am
Munir wrote: Thu May 07, 2020 10:52 am The graphs stop at age 65. Why?
terran wrote: Thu May 07, 2020 10:29 am Interesting. I'd be interested how this could be adapted to early retirees: steeper/higher savings rates, and shorter accumulation periods. Is there anything more you can share that would help one figure out the optimal path under such altered conditions?
I hinted at this earlier, but saving for retirement (accumulation) and spending in retirement (decumulation) are very different problems.

Accumulation is essentially an investment "problem" but decumulation isn't: it's an income "problem", which requires a distinct approach that was beyond the scope of the OP.

This new rule of thumb is designed for people who are turning income into wealth. People who are turning wealth into income (i.e. retirees) probably shouldn't be relying on asset allocation as their primary method of control.
You're right, I probably should have said "people planning to retire early", rather than "early retirees." I feel pretty comfortable with the information out there on the rising equity glidepath for decumulation, but I haven't seen as much about accumulation, so I'd be curious how this might be adapted to accumulation for people who plan to retire early. For example, 2*(40-40) = 0% bonds probably doesn't make sense for someone retiring in 5 years at age 45, but just shifting that persons "age" to normalize for a 65 year old retirement with 2*(60-40) = 40% bonds might not make sense for someone saving at a much higher rate than the 15% this model is made for. I'm wondering how you arrived at the optimal glidepath and how it might be adjusted for a higher savings rate and shorter accumulation period? My guess would be that the glidepath should get steeper as the savings rate increases.
vitaflo wrote: Thu May 07, 2020 11:00 am An even better rule of thumb is not to use age at all. Age is just a proxy for portfolio size, specifically how much you have in your portfolio vs what you need in your portfolio to retire on.

So just use that instead. The math is more complicated, but it means your glide path will reach your desired AA regardless of when you retire, which is especially useful for those who want to retire early.
That could be an interesting solution to my question above, what might that look like?
User avatar
Topic Author
vineviz
Posts: 9933
Joined: Tue May 15, 2018 1:55 pm

Re: A bond rule of thumb better than "age in bonds"

Post by vineviz »

HomerJ wrote: Thu May 07, 2020 12:55 pm What the heck is "recklessly conservative"?
"Age in bonds" is recklessly conservative.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch
petulant
Posts: 1901
Joined: Thu Sep 22, 2016 1:09 pm

Re: A bond rule of thumb better than "age in bonds"

Post by petulant »

But can you make a quadratic equation that also incorporates the rising equity glidepath that some people recommend for after retirement? :twisted:
User avatar
goodenyou
Posts: 2674
Joined: Sun Jan 31, 2010 11:57 pm
Location: Skating to Where the Puck is Going to Be..or on the golf course

Re: A bond rule of thumb better than "age in bonds"

Post by goodenyou »

fatcoffeedrinker wrote: Thu May 07, 2020 12:31 pm
willthrill81 wrote: Thu May 07, 2020 12:07 pm
midareff wrote: Thu May 07, 2020 10:39 am Bonds = 2 * (age - 40) Let me see..... age 72, minus 40 = 32*2 = 64% bonds... it doesn't work for me.
Vineviz's formula is probably best suited for accumulators, not those in the withdrawal phase. Presumably, you would only use the formula to increase your bond allocation until it reached the AA you wanted to at least begin your retirement at.
Even for accumlators, I think at least two other factors come into play other than age: years until target retirement date and need to take risk. I'm 51 and plan to retire in 3 or so years. I have been very fortunate lately in that my comp has increased to the point where I can now save 5x expenses per year. So for me, I don't want a market crash to push out my retirement date, and because of my savings rate, I don't need much return to hit my retirement goal. That is why I am 40/60 right now. Super conservative for 51, yes, but it works for my situation at this time given my goals. After I retire, I will likely ramp the stock allocation up over time, as I don't want to be 40/60 forever. Maybe end at 60/40 or 65/35, or maybe just keep 10x expenses in FI and invest the rest in stocks -- not quite sure yet.
For someone who can save their way to the finish line, it would be foolish to pick an AA that could derail their plans. Keeping up with inflation may be the primary goal. The challenge for some is whether they will feel foolish that they left a significant portion "on the table" even with minimum of risk.
"Ignorance more frequently begets confidence than does knowledge" | “Do you know how to make a rain dance work? Dance until it rains”
User avatar
goodenyou
Posts: 2674
Joined: Sun Jan 31, 2010 11:57 pm
Location: Skating to Where the Puck is Going to Be..or on the golf course

Re: A bond rule of thumb better than "age in bonds"

Post by goodenyou »

vineviz wrote: Thu May 07, 2020 1:06 pm
HomerJ wrote: Thu May 07, 2020 12:55 pm What the heck is "recklessly conservative"?
"Age in bonds" is recklessly conservative.
Except when you are 80.
"Ignorance more frequently begets confidence than does knowledge" | “Do you know how to make a rain dance work? Dance until it rains”
Wanderingwheelz
Posts: 1185
Joined: Mon Mar 04, 2019 9:52 am

Re: A bond rule of thumb better than "age in bonds"

Post by Wanderingwheelz »

If you’ve “made it” and don’t want to risk losing it, it suggest not going beyond the classic 60/40, or maybe 50/30/10 if you want to add some gold.

Age has very little to do with whether or not a person has “made it”, so ages based rules are of little use to a person who has enough money that they won’t run out.

I’m comfortable with a 50% stock allocation, not counting my real estate equity. At 48 many here would think that’s far too conservative, but we are all wired differently. I’d way rather cut spending at 70 if the world goes sideways, that have to go work in an Amazon warehouse.

Inflation plus 1% is perfectly fine with me, since I have enough money and the flexibility to adjust my lifestyle if having half my assets in stocks turns out to be too conservative over the long-haul.
3 Fund Portfolio. 70/30 AA. No mortgage. Simplicity is key.
User avatar
SimpleGift
Posts: 3869
Joined: Tue Feb 08, 2011 3:45 pm
Location: Central Oregon

Re: A bond rule of thumb better than "age in bonds"

Post by SimpleGift »

fatcoffeedrinker wrote: Thu May 07, 2020 12:42 pm Interesting. Rough formula based on that graph is bonds = (5*age - 40)/6.
The table below shows the actual allocation of the "average Boglehead" vs. their age in bonds from the 2015 poll:
But as the graph upthread indicates, the dispersion from the trend line is so great that the "average allocation" is fairly meaningless. Plus, the poll was taken 5 years ago, so Bogleheads' allocations may well have shifted a bit since then.
Triple digit golfer
Posts: 7390
Joined: Mon May 18, 2009 5:57 pm

Re: A bond rule of thumb better than "age in bonds"

Post by Triple digit golfer »

vineviz wrote: Thu May 07, 2020 9:58 am For those who prefer to think in terms of wealth, rather than age, there's a roughly equivalent wealth-based rule of thumb.

Compute your expected retirement withdrawals (in current dollars) as a multiple of your current retirement savings: this is (in percentage terms) your bond allocation.

So, for example, imagine that your current retirement savaging are $200,000 and you expect your annual portfolio withdrawals in retirement will be $20,000. 200,000 divided by 20,000 is 10. Your bond allocation is 10%.

If your current retirement savaging are $800,000 and you expect your annual portfolio withdrawals in retirement will be $20,000 then your bond allocation is 40% ( 800,000 divided by 20,000 is 40).

This will produce a slightly different glide path than the age-based rule, but the shape of the glide path will be roughly similar for most investors.
Why would someone with 25x expenses want to be 75/25? That is way too aggressive for my liking. Imagine a 60 year-old freshly retired person with just 25x in expenses. Why be anything more than 50/50, which should be enough for a 30 year retirement anyway? A poor sequence of returns at 75/25 could derail his retirement.
Nowizard
Posts: 3309
Joined: Tue Oct 23, 2007 5:33 pm

Re: A bond rule of thumb better than "age in bonds"

Post by Nowizard »

Far too many variables to have a formula except for those who are attracted to them. Individual circumstances are crucial. As one example, if someone has an inflation indexed pension and SS that meets, say, 80% of annual expenses of $120,000, vs. someone who has SS that meets 30% of the same expense annually, there may be marked differences in bond allocations.

Tim
User avatar
Topic Author
vineviz
Posts: 9933
Joined: Tue May 15, 2018 1:55 pm

Re: A bond rule of thumb better than "age in bonds"

Post by vineviz »

Triple digit golfer wrote: Thu May 07, 2020 1:30 pm Why would someone with 25x expenses want to be 75/25? That is way too aggressive for my liking. Imagine a 60 year-old freshly retired person with just 25x in expenses. Why be anything more than 50/50, which should be enough for a 30 year retirement anyway? A poor sequence of returns at 75/25 could derail his retirement.
As csmath said upthread, the role of a rule of thumb is get you in the vicinity of a good solution. I hope readers will take my suggestion as it was intended (i.e. "this is better than age in bonds) rather than imagining it as substitute for actual planning.

If you're a 60-year old investor with 25x expenses who thinks 75/25 is too aggressive then by all means choose something more conservative.

Economically, though, 75/25 is definitely near - if not on - the part of the allocation curve that will maximize SWR (and minimize sequence of return risk). 75/25 certainly gives the retiree (who, actuarially, probably should be planning for a 40-year retirement anyway) a better shot at making it through a long retirement than 50/50 would.
Last edited by vineviz on Thu May 07, 2020 2:40 pm, edited 1 time in total.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch
User avatar
Topic Author
vineviz
Posts: 9933
Joined: Tue May 15, 2018 1:55 pm

Re: A bond rule of thumb better than "age in bonds"

Post by vineviz »

terran wrote: Thu May 07, 2020 1:04 pm That could be an interesting solution to my question above, what might that look like?

It think the best approach for people planning retirements much earlier than assumed here, and people with savings rates much higher than the rates I assumed, should probably be doing an explicit NPV (or lifecycle) calculation to arrive at their asset allocation strategy.

Ayres and Nalebuff outlined the approach in their "Lifecycle Investing" book, and I took the liberty of adapting their spreadsheet for Google Sheets: link
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch
hudson
Posts: 3890
Joined: Fri Apr 06, 2007 9:15 am

Re: A bond rule of thumb better than "age in bonds"

Post by hudson »

vineviz wrote: Thu May 07, 2020 9:40 am I've been trying to come up with an easily computable rule that investors can use to set their own glide path, and I think this will do it.
Bonds = 2 * (age - 40)
I think that both the vineviz's formula or the age in bonds formula will work for a beginning investor. I think those formulas are just starting points in the decision that a beginning investor needs to make. Those of us who have been doing this for a while do not need a rule of thumb; we know exactly what percentage of bonds works for us.
The beginning investor should take Vineviz's formula or age in bonds in mind and then read the Wiki, Swedroe, W. Bernstein, Larimore, Ferri and others. When they finish, they should know more about what fits them.

Nice graph vineviz! I didn't want to do the math, your graph helped me understand.
Last edited by hudson on Thu May 07, 2020 4:53 pm, edited 1 time in total.
HawkeyePierce
Posts: 1828
Joined: Tue Mar 05, 2019 10:29 pm
Location: Colorado

Re: A bond rule of thumb better than "age in bonds"

Post by HawkeyePierce »

I strongly disagree that "age in bonds" is recklessly conservative. Two reasons:

1) You can't out-invest an insufficient savings rate. If 60/40 won't get you to your goals, 100/0 probably won't either. Going back to 1972, the difference in returns between 60/40 and 100/0 is not significant compared to the return on increasing your savings rate.

2) I firmly believe the average investor is less likely to panic-sell at 60/40 than 100/0. It may be conservative, but not recklessly so if it helps the average investor stay the course.

Personally, I think a 60/40 target allocation or balanced fund is perfect for 99% of investors by minimizing regret and maximizing the chance of success (where success is defined as "able to retire without going broke").
Post Reply