Watch out for hidden risk tolerance assumptions in SWR claims

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
User avatar
Topic Author
nisiprius
Advisory Board
Posts: 47253
Joined: Thu Jul 26, 2007 9:33 am
Location: The terrestrial, globular, planetary hunk of matter, flattened at the poles, is my abode.--O. Henry

Watch out for hidden risk tolerance assumptions in SWR claims

Post by nisiprius »

(Background: SWR = "safe" (or "sustainable") "withdrawal rate," how much can a retiree withdraw annually from an investment portfolio which fluctuates in value, with little risk of running out of money. The "4% rule" is a common answer. The percentage is a percentage of the initial portfolio value. Once chosen, the withdrawals are adjusted for inflation, but not adjusted for investment performance).

I just want to illustrate a fairly obvious point. I don't believe it's controversial, but I think it's important and not often presented. It's worth making because of the attention that gets paid to SWR as a figure of merit for portfolios.

I'm going to use the Vanguard Nest Egg Calculator to illustrate the point, but it's not sensitive to the tool chosen--I "discovered" it myself years ago using the Fidelity Retirement Income planner. Because it is truly a Monte Carlo simulation, re-runs do not produce strictly identical results. In all cases I will set tool's parameters for 30 years, $1 million, and 0% cash, and vary withdrawal rates and stock/bond allocation. I'm going to use extreme values, but the same qualitative effects are observed for more reasonable values.

We'll start by assuming a 7% withdrawal rate, which is obviously too high--but not a complete straw man because Peter Lynch actually suggested in 1995 that a 100% stocks portfolio could sustain 7%.

With 25/75, the portfolio only survived 10% of the time.
With 75/25, 47%.
With 100/0, Peter Lynch's suggestion, 53%.

Image
Image
Image

Notice two points:
  • It is perfectly true that higher stock allocations led to higher success rates, but...
  • Even at 100% stocks, the success rate was still only 53%.
Now, we will try it assuming a 3.5% withdrawal rate. At this withdrawal rate,

With 25/75, the portfolio survived 96% of the time.
With 75/25, 94%. Slightly less.
With 100/0, 91%.

Image
Image
Image
  • At conservative withdrawal rates, the success rate was not very sensitive to the allocation choice.
  • In every study I've seen, for sensible success rates the result was always the same: the success rates and/or SWR with a chosen success rate was almost flat over, let's say 25/75 to 75/25, and drooped at both ends.
  • 100% stocks is unwise. 100% bonds is even worse. But anything within 25/75 to 75/25 is OK from an SWR standpoint.
To show that is correct assuming a conservative withdrawal rate, I reproduce the results of a study by William Bengen, which holds success rate constant and shows how SWR varies. My point is not the absolute numbers, but the flatness of the center part of the curve.

Source

Image

The point I want to make is to be critical of claims that some portfolio change "increases the safe withdrawal rate," because whether or not it does may depend on a smuggled-in assumption about the investor's risk tolerance. All such assessments should be made on the basis of a targeted failure rate, and the choice of that target depends on the risk tolerance of the investor.

Our crude exploration showed that an investor able to tolerate a 50% chance of running out of money is best served by a high stock allocation--100% stocks, in fact; but an investor only able to tolerate a 5% chance is almost indifferent to the allocation decision and is better served by 25/75 or 75/25 or 100/0.

(Or by 0/100. The success rates in Vanguard's simulation were 1% for 0/100 at a 7% withdrawal--1%! And only 86% at a 3.5% withdrawal).

Unfortunately, many SWR studies simply assume without any discussion that a 5% failure rate is appropriate.

Is it? That's a difficult question. Simply on the face of it, and as a planning assumption, if you were handed a pair of dice at the start of retirement, and told to roll the dice and accept that your chosen plan would fail if you rolled a twelve, would you accept that as a sound plan? That is not the only way to frame the decision, but it's relevant.
Last edited by nisiprius on Thu Aug 04, 2022 7:44 am, edited 2 times in total.
Annual income twenty pounds, annual expenditure nineteen nineteen and six, result happiness; Annual income twenty pounds, annual expenditure twenty pounds ought and six, result misery.
User avatar
burritoLover
Posts: 2550
Joined: Sun Jul 05, 2020 12:13 pm

Re: "Stocks increase SWR" assumes a risk-tolerant investor

Post by burritoLover »

The following chart gets thrown around a lot but shows much different results:

Image
User avatar
Topic Author
nisiprius
Advisory Board
Posts: 47253
Joined: Thu Jul 26, 2007 9:33 am
Location: The terrestrial, globular, planetary hunk of matter, flattened at the poles, is my abode.--O. Henry

Re: "Stocks increase SWR" assumes a risk-tolerant investor

Post by nisiprius »

It isn't "very different results." It actually illustrates my point. (There is a smuggled-in risk tolerance assumption in the choice of color coding). At lower withdrawal rates, everything is green over a broad range, all the way down to 25% stocks at and 40 years at least. The effect of "more stocks improves survival rates" is only true at the higher withdrawal rates, where nothing is very green.

The reason is pretty simple. If you choose a withdrawal rate that bonds can't sustain, then stocks will give you a shot at sustaining it--but not much certainty.
Last edited by nisiprius on Thu Aug 04, 2022 7:23 am, edited 1 time in total.
Annual income twenty pounds, annual expenditure nineteen nineteen and six, result happiness; Annual income twenty pounds, annual expenditure twenty pounds ought and six, result misery.
muffins14
Posts: 2436
Joined: Wed Oct 26, 2016 4:14 am
Location: New York

Re: "Stocks increase SWR" assumes a risk-tolerant investor

Post by muffins14 »

I’d recommend removing everything you have on the 7% withdrawal rate to make the post more concise. No one here is going to plan for 7%, so it’s a bit of a distraction
35% VTI, 25% AVUV, 15% IXUS, 15% AVDV, 10% VWO
User avatar
Topic Author
nisiprius
Advisory Board
Posts: 47253
Joined: Thu Jul 26, 2007 9:33 am
Location: The terrestrial, globular, planetary hunk of matter, flattened at the poles, is my abode.--O. Henry

Re: "Stocks increase SWR" assumes a risk-tolerant investor

Post by nisiprius »

muffins14 wrote: Thu Aug 04, 2022 7:23 am I’d recommend removing everything you have on the 7% withdrawal rate to make the post more concise. No one here is going to plan for 7%, so it’s a bit of a distraction
My point is that stocks make a huge increase in SWR for a very-risk-tolerant investor, and very little difference for a risk-averse investor. What you see is going to depend on what you are assuming for risk tolerance.

The reason for choosing an unrealistic 7% is to exhibit a very clear case where it is technically true that a higher stock allocation "improved" portfolio survival... but that the best survival percentage is unacceptably low for most people.

If you go down into a sensible range, then the argument deteriorates to "how do you decide, for planning purposes, what number to use as an acceptable success/failure rate?" and becomes a meaningless exercise in "what is the objectively correct amount of risk aversion?" Is 95% portfolio survival meaningfully different from 90%? (Which is just another form of the question "is an 8% standard deviation meaningfully different from 9%?")
Last edited by nisiprius on Thu Aug 04, 2022 7:37 am, edited 1 time in total.
Annual income twenty pounds, annual expenditure nineteen nineteen and six, result happiness; Annual income twenty pounds, annual expenditure twenty pounds ought and six, result misery.
petulant
Posts: 2225
Joined: Thu Sep 22, 2016 1:09 pm

Re: "Stocks increase SWR" assumes a risk-tolerant investor

Post by petulant »

While I agree with the underlying theory, my impression is that people talking about "improving SWR" actually do hold a constant, conservative risk tolerance for discussion. The oft-unstated metric is a 0% failure rate for 30 years, which is very conservative. For example, that is the simba backtesting spreadsheet's definition of a safe withdrawal rate. A mix of assets that improves the SWR by 25 or 50 bps on that metric is meaningful.
User avatar
burritoLover
Posts: 2550
Joined: Sun Jul 05, 2020 12:13 pm

Re: "Stocks increase SWR" assumes a risk-tolerant investor

Post by burritoLover »

nisiprius wrote: Thu Aug 04, 2022 7:21 am It isn't "very different results." It actually illustrates my point. (There is a smuggled-in risk tolerance assumption in the choice of color coding). At lower withdrawal rates, everything is green over a broad range, all the way down to 25% stocks at and 40 years at least. The effect of "more stocks improves survival rates" is only true at the higher withdrawal rates, where nothing is very green.

The reason is pretty simple. If you choose a withdrawal rate that bonds can't sustain, then stocks will give you a shot at sustaining it--but not much certainty.
Ah, I saw your graph go to 50 years, but you were using the 30 year results from it.
User avatar
Topic Author
nisiprius
Advisory Board
Posts: 47253
Joined: Thu Jul 26, 2007 9:33 am
Location: The terrestrial, globular, planetary hunk of matter, flattened at the poles, is my abode.--O. Henry

Re: "Stocks increase SWR" assumes a risk-tolerant investor

Post by nisiprius »

petulant wrote: Thu Aug 04, 2022 7:37 am While I agree with the underlying theory, my impression is that people talking about "improving SWR" actually do hold a constant, conservative risk tolerance for discussion. The oft-unstated metric is a 0% failure rate for 30 years, which is very conservative. For example, that is the simba backtesting spreadsheet's definition of a safe withdrawal rate. A mix of assets that improves the SWR by 25 or 50 bps on that metric is meaningful.
OK, point taken.

I hope people won't be too annoyed if I respond to the criticism by softening the thread title.

It was "Stocks increase SWR" assumes a risk-tolerant investor". I'm going to change it to "Watch out for hidden risk tolerance assumptions in SWR claims."
Last edited by nisiprius on Thu Aug 04, 2022 7:46 am, edited 2 times in total.
Annual income twenty pounds, annual expenditure nineteen nineteen and six, result happiness; Annual income twenty pounds, annual expenditure twenty pounds ought and six, result misery.
Tamalak
Posts: 1399
Joined: Fri May 06, 2016 2:29 pm

Re: "Stocks increase SWR" assumes a risk-tolerant investor

Post by Tamalak »

It's a good post. I'm risk-tolerant so far - I didn't even flinch at the covid crash or this year's slump - but this is the accumulation phase. During accumulation phase bears are, in some cases, good, and even if they're bad all they do is shift your retirement schedule a bit.

I sense retirement will be more of a dependent, helpless feeling, since I'm not adding anything more to my portfolio. What will a covid lightning strike feel like then?
Kevin K
Posts: 281
Joined: Sun Aug 26, 2007 7:47 pm

Re: Watch out for hidden risk tolerance assumptions in SWR claims

Post by Kevin K »

Tyler @ Portfolio Charts has a nifty set of tools for looking at SWR's along with several articles that taken together are far and away the best updating of Bengen's work I've seen. In addition to being able to model SWR and PWR (Perpetual Withdrawal Rate) for any allocation he also provides those numbers for a large number of popular "lazy" portfolios, including Boglehead's classics like the 60:40 and Three Fund.

https://portfoliocharts.com/portfolio/withdrawal-rates/
User avatar
vineviz
Posts: 12309
Joined: Tue May 15, 2018 1:55 pm
Location: Baltimore, MD

Re: "Stocks increase SWR" assumes a risk-tolerant investor

Post by vineviz »

petulant wrote: Thu Aug 04, 2022 7:37 am While I agree with the underlying theory, my impression is that people talking about "improving SWR" actually do hold a constant, conservative risk tolerance for discussion. The oft-unstated metric is a 0% failure rate for 30 years, which is very conservative. For example, that is the simba backtesting spreadsheet's definition of a safe withdrawal rate. A mix of assets that improves the SWR by 25 or 50 bps on that metric is meaningful.
Probabilities are hard.

Many backward-looking SWR studies gloss over the fact that the worst REALIZED sustainable withdrawal rate was not actually the worst POSSIBLE rate that historic conditions could have produced.

In other words, the worst historical SWR was (based on my analysis) roughly equivalent to a 5% or 10% probability outcome using the kind of Monte Carlo simulations employed by nisiprius in the OP.

In reality, planning for a truly 0% failure rate using forward-looking estimates would be insanely conservative whereas planning for a 0% failure rate using backward-looking outcomes is probably more aggressive than most people understand.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch
Escapevelocity
Posts: 971
Joined: Mon Feb 18, 2019 8:32 am

Re: Watch out for hidden risk tolerance assumptions in SWR claims

Post by Escapevelocity »

nisiprius wrote: Thu Aug 04, 2022 7:00 am Unfortunately, many SWR studies simply assume without any discussion that a 5% failure rate is appropriate.

Is it? That's a difficult question. Simply on the face of it, and as a planning assumption, if you were handed a pair of dice at the start of retirement, and told to roll the dice and accept that your chosen plan would fail if you rolled a twelve, would you accept that as a sound plan? That is not the only way to frame the decision, but it's relevant.
On this point, it very much depends on the role of the portfolio withdrawals in terms of funding expenses in retirement. If social security and pensions (or other reliable income streams) are covering a large percentage of essential expenses as is very common, then it is quite reasonable to take a 5% risk of failure on the portfolio failing in the tail end of a retirement when discretionary spending is likely to fall anyway and when there is also a high probability of the retiree being expired at that stage as well.
Tamalak
Posts: 1399
Joined: Fri May 06, 2016 2:29 pm

Re: Watch out for hidden risk tolerance assumptions in SWR claims

Post by Tamalak »

What I don't like about these "odds your money will last" sites is they seem to assume that you're equally likely to retire every year.

But people are much more likely to hit their number when the market was high. Nobody hit it at the bottom of 2008. So I feel like it might be overly optimistic.
martincmartin
Posts: 473
Joined: Wed Jul 02, 2014 3:04 pm

Re: Watch out for hidden risk tolerance assumptions in SWR claims

Post by martincmartin »

Another confusing point is that the word "risk" has multiple definitions. For a financial advisor, and most writing, it means variation of an investment, as measured by the standard deviation of, say, daily returns. This is the sort of thing that causes people to be surprised that their retirement accounts sometimes go down, and post in our forums, and perhaps panic sell on the way down. So it's a valid and useful definition of "risk", I just which it wasn't considered the default definition of risk.

Another definition of risk, in a SWR context, is "chance of running out of money in retirement." In this case, if you retire with 25x annual expenses and make the standard assumptions (30 year retirement, ignore taxes, no SS, expenses increase exactly with CPI-U, etc.), then 100% bonds is riskier than stocks. For example, in Turkey right now, inflation is about 80% while interest rates are low. I don't know what's happening to Turkish stocks, but bonds are quickly losing purchasing power.

Anyway, "risk tolerance" usually means ability not to panic sell on a big drop. Are you using it in that sense, or some other way?
dorster
Posts: 74
Joined: Tue Aug 15, 2017 4:20 pm

Re: Watch out for hidden risk tolerance assumptions in SWR claims

Post by dorster »

nisiprius wrote: Thu Aug 04, 2022 7:00 am The point I want to make is to be critical of claims that some portfolio change "increases the safe withdrawal rate,"
Your post reminded me of a great post on EREVN's blog on medium. How important is asset allocation versus withdrawal rates in retirement? Their blog has been inactive for a couple years but I believe they are a frequent poster on this forum.

Here are the concluding two paragraphs:
EREVN wrote:Instead of stressing about trying to pick “the right” asset allocation, you’re better off picking anything reasonable and ignoring every other asset allocation internet discussion for the rest of your life…and then working an extra six or twelve months to pad out your retirement fund before retiring.

And, as a corollary, if asset allocation is this unimportant…all of the other, even more esoteric decisions (including ones I’ve written about…!) are even less important.
EREVN's blog has a lot of great posts and I think this one illustrates your point (or what I think is your point) in an interesting and novel way.
petulant
Posts: 2225
Joined: Thu Sep 22, 2016 1:09 pm

Re: Watch out for hidden risk tolerance assumptions in SWR claims

Post by petulant »

dorster wrote: Thu Aug 04, 2022 9:19 am
nisiprius wrote: Thu Aug 04, 2022 7:00 am The point I want to make is to be critical of claims that some portfolio change "increases the safe withdrawal rate,"
Your post reminded me of a great post on EREVN's blog on medium. How important is asset allocation versus withdrawal rates in retirement? Their blog has been inactive for a couple years but I believe they are a frequent poster on this forum.

Here are the concluding two paragraphs:
EREVN wrote:Instead of stressing about trying to pick “the right” asset allocation, you’re better off picking anything reasonable and ignoring every other asset allocation internet discussion for the rest of your life…and then working an extra six or twelve months to pad out your retirement fund before retiring.

And, as a corollary, if asset allocation is this unimportant…all of the other, even more esoteric decisions (including ones I’ve written about…!) are even less important.
EREVN's blog has a lot of great posts and I think this one illustrates your point (or what I think is your point) in an interesting and novel way.
Welp now I've got about 20 tabs open to read during breaks...
sailaway
Posts: 5485
Joined: Fri May 12, 2017 1:11 pm

Re: Watch out for hidden risk tolerance assumptions in SWR claims

Post by sailaway »

Fat fingered quoting removed.

We are going by the "pick an AA and pad the numbers a bit" method.
59Gibson
Posts: 1386
Joined: Mon Dec 07, 2020 8:41 am

Re: Watch out for hidden risk tolerance assumptions in SWR claims

Post by 59Gibson »

Concerning your 5% failure example..it doesn't mean the retiree goes broke. It simply means an adjustment is required. Depending on one's needs/ spending, failure may be too strong of a word imo.
I am a proponent for 25-75% equity
Last edited by 59Gibson on Thu Aug 04, 2022 9:55 am, edited 2 times in total.
User avatar
willthrill81
Posts: 32250
Joined: Thu Jan 26, 2017 3:17 pm
Location: USA

Re: "Stocks increase SWR" assumes a risk-tolerant investor

Post by willthrill81 »

nisiprius wrote: Thu Aug 04, 2022 7:32 amMy point is that stocks make a huge increase in SWR for a very-risk-tolerant investor, and very little difference for a risk-averse investor.
I think it's worth pointing out that your statement illustrates a very common problem with the terms 'risk tolerant' and 'risk averse' as we commonly refer to them.

The investor with the higher tolerance for 'risk' had an objectively lower likelihood of running out of money than the investor with the low tolerance for 'risk'.

In a very real way, being more tolerant of 'risk' actually resulted in more safety in terms of achieving the desired outcome.

This demonstrates very well that what we often refer to as risk tolerance is actually volatility tolerance. The two are not the same.
I have left the forum but occasionally check PMs.
JackoC
Posts: 3636
Joined: Sun Aug 12, 2018 11:14 am

Re: "Stocks increase SWR" assumes a risk-tolerant investor

Post by JackoC »

vineviz wrote: Thu Aug 04, 2022 8:53 am
petulant wrote: Thu Aug 04, 2022 7:37 am While I agree with the underlying theory, my impression is that people talking about "improving SWR" actually do hold a constant, conservative risk tolerance for discussion. The oft-unstated metric is a 0% failure rate for 30 years, which is very conservative. For example, that is the simba backtesting spreadsheet's definition of a safe withdrawal rate. A mix of assets that improves the SWR by 25 or 50 bps on that metric is meaningful.
Probabilities are hard.

Many backward-looking SWR studies gloss over the fact that the worst REALIZED sustainable withdrawal rate was not actually the worst POSSIBLE rate that historic conditions could have produced.

In other words, the worst historical SWR was (based on my analysis) roughly equivalent to a 5% or 10% probability outcome using the kind of Monte Carlo simulations employed by nisiprius in the OP.

In reality, planning for a truly 0% failure rate using forward-looking estimates would be insanely conservative whereas planning for a 0% failure rate using backward-looking outcomes is probably more aggressive than most people understand.
I agree, the additional complication is how literally to take a 'probability' derived from a given method. The Vanguard simulation and say FireCalc both use the past distribution of returns (shorter one in Vanguard's case* but not the main issue). As you suggest Firecalc uses randomly selected historical paths for the say 30 yrs. The Vanguard method randomly selects single years and pieces them together into 30 yr sequences. That gives worse sequences that 'could have happened' in the past distribution than the ones which actually did. But there's more... :happy (which you well know but to clarify in general). In both cases the overall center of gravity of the return distribution is that of the historical. But we're starting now from a point of especially high risk asset valuation and low riskless yield, with no assurance (no basis at all IMO to assume) 'returns are more likely to average out in a long period from now to whatever they did in 1871-now' (or Vanguard's dates to now). If forced to determine my asset allocation/plan with a simulation tool I'd want the center of the distribution to be *now's* expected return. I might use the historical variance and higher moments (skew etc) for lack of a better alternative. But what showed up failing 5% of the time in Firecalc would generally fail more often in Vanguard's tool I believe, and definitely lot more in my hypothetical tool with the center of the distribution 2-3% points pa lower than historical. In any case you can't directly compare a 'failure %' from a tool with inevitably controversial assumptions (not just 'nobody knows the future' but that others would do the calculation a pretty different way) to accepting those odds in a game of (known, fair) dice with the same stakes.

That said the exercise IMO is a good one generally reinforcing the intuition of the Graham Rule (25>75 risk assets is the range of wise allocation) and I don't think a different simulation technique would change that. And it doesn't disagree with the simpler means by which I've chosen an allocation within that range (in retirement). It also seems intuitive that a person willing to accept a relatively high chance of failure would be the most likely to adopt a risk asset allocation stretching or exceeding that range.

*"For stock market returns we use the Standard & Poor's 500 Index from 1926 to 1970, the Dow Jones Wilshire 5000 Index from 1971 through April 2005, and the MSCI US Broad Market Index thereafter. For bond market returns, we use the Standard & Poor's High Grade Corporate Index from 1926 to 1968, the Citigroup High Grade Index from 1969 to 1972, the Barclays US Long Credit AA Index from 1973-1975, and the Barclays Capital US Aggregate Bond Index thereafter. For the returns on short-term reserves (i.e., "cash"), we use the Citigroup 3-Month Treasury Bill Index. For inflation, we use the changes in the annual Consumer Price Index from 1926 through last year."
User avatar
willthrill81
Posts: 32250
Joined: Thu Jan 26, 2017 3:17 pm
Location: USA

Re: "Stocks increase SWR" assumes a risk-tolerant investor

Post by willthrill81 »

vineviz wrote: Thu Aug 04, 2022 8:53 am In reality, planning for a truly 0% failure rate using forward-looking estimates would be insanely conservative whereas planning for a 0% failure rate using backward-looking outcomes is probably more aggressive than most people understand.
It's literally impossible to have a 0% failure rate because doing so would require knowledge of the future.

Statistics can never be used to indicate that there is literally a 0% probability of any event occurring. At best, they can only indicate that a given event may be extremely unlikely to occur. But they can never indicate a 0% probability.

Bernstein has argued that non-portfolio risks are high enough that the best probability of a 30 year retirement succeeding in the sense of no adjustments being necessary is about 80%.
I have left the forum but occasionally check PMs.
petulant
Posts: 2225
Joined: Thu Sep 22, 2016 1:09 pm

Re: "Stocks increase SWR" assumes a risk-tolerant investor

Post by petulant »

The people who look at and discuss a historical SWR defined as 0% failure over 30 years (including user willthrill81) are aware of the difference pointed out by vineviz, and they don't assume that 0% failure under such analysis actually means 0% probability of failure in real life (as revealed by willthrill81's post). But vineviz was careful to phrase his post that "most people" don't understand the difference, which might be true. It seems to me that SWR studies can provide meaningful information that astute readers can understand but which many people would misinterpret.
User avatar
willthrill81
Posts: 32250
Joined: Thu Jan 26, 2017 3:17 pm
Location: USA

Re: "Stocks increase SWR" assumes a risk-tolerant investor

Post by willthrill81 »

petulant wrote: Thu Aug 04, 2022 10:08 amIt seems to me that SWR studies can provide meaningful information that astute readers can understand but which many people would misinterpret.
Absolutely. Look at what the MMM crowd and many in the FIRE community have done with the '4% rule of thumb'. They've erroneously concluded that it applies to retirements of any given length, despite Bengen explicitly saying in his seminal paper that longer retirements than 30 years would have resulted in a lower SWR than 4%.

We see a recurring problem over and over across innumerable topics, not just personal finance: humans crave simple, black-and-white answers, even though reality is detailed and usually very nuanced.
I have left the forum but occasionally check PMs.
User avatar
snackdog
Posts: 1658
Joined: Wed Nov 12, 2014 4:57 am

Re: Watch out for hidden risk tolerance assumptions in SWR claims

Post by snackdog »

What is “hidden”? Isn’t it intuitive that stocks are more volatile than bonds and there better suited to a higher and riskier withdrawal profile?
User avatar
vineviz
Posts: 12309
Joined: Tue May 15, 2018 1:55 pm
Location: Baltimore, MD

Re: "Stocks increase SWR" assumes a risk-tolerant investor

Post by vineviz »

willthrill81 wrote: Thu Aug 04, 2022 10:02 am Statistics can never be used to indicate that there is literally a 0% probability of any event occurring.
There's an oppositional streak in me that is dying to point out something like "well, it depends on how many decimal points you put after the zero in 0%".

But I'm above that kind of snark, thankfully. ;)
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch
Tamalak
Posts: 1399
Joined: Fri May 06, 2016 2:29 pm

Re: Watch out for hidden risk tolerance assumptions in SWR claims

Post by Tamalak »

snackdog wrote: Thu Aug 04, 2022 10:22 am What is “hidden”? Isn’t it intuitive that stocks are more volatile than bonds and there better suited to a higher and riskier withdrawal profile?
It's hidden by the numbers, like that colored chart above. The more stocks you have the more green (positive chance of success) you get. No brainer to have 100% stocks, right?
User avatar
rocket354
Posts: 531
Joined: Mon Dec 14, 2015 12:31 pm

Re: Watch out for hidden risk tolerance assumptions in SWR claims

Post by rocket354 »

My issue with "fail or not" analysis is that there is a huge middle ground that could be considered a failure that is not taken into account. If I'm left with $2.00 after 30 years with a given WR and AA...is that a success if I'm still alive?

One thing I noticed is that with higher equity allocations you get a slight bump in the failure rate in the mid-WR ranges (3.2-4.0%) but you also see a decrease, particularly in longer time frames (40+ years), in the "middling" outcomes--which I would cursorily define as having an inflation-adjusted balance less than the starting balance. These outcomes in many cases can be considered failures because one can be left stressfully withdrawing from a dwindling portfolio that you might actually outlive, and therefore can cause effective failures where one returns to work or decides to make dramatic changes to one's lifestyle to survive (the "eating cat food" outcome).

Ultimately, the decision is not, in my eyes, between:

90% success/10% failure
95% success/5% failure

It is potentially between:

85% success/5% middling/10% failure
75% success/20% middling/5% failure

Note that these numbers are made up for illustrative purposes. However, I think one can see that the decision in terms of AA and WR becomes murkier when the third outcome is introduced.
User avatar
Riprap
Posts: 657
Joined: Thu Jan 29, 2009 2:08 pm

Re: Watch out for hidden risk tolerance assumptions in SWR claims

Post by Riprap »

I guess ole Ben Graham was onto something when he said the defensive investor "should NEVER have less than 25% or more than 75% of his funds in common stock." I've always wondered how he came up with that range? I'm pretty sure he didn't have access to computerized numerical methods that are common today.

Side question: How are the graphics posted to this thread? They are fantastic, a picture being worth a thousand words. I've never been able to figure it out.
JackoC
Posts: 3636
Joined: Sun Aug 12, 2018 11:14 am

Re: "Stocks increase SWR" assumes a risk-tolerant investor

Post by JackoC »

willthrill81 wrote: Thu Aug 04, 2022 10:02 am
vineviz wrote: Thu Aug 04, 2022 8:53 am In reality, planning for a truly 0% failure rate using forward-looking estimates would be insanely conservative whereas planning for a 0% failure rate using backward-looking outcomes is probably more aggressive than most people understand.
It's literally impossible to have a 0% failure rate because doing so would require knowledge of the future.

Statistics can never be used to indicate that there is literally a 0% probability of any event occurring. At best, they can only indicate that a given event may be extremely unlikely to occur. But they can never indicate a 0% probability.

Bernstein has argued that non-portfolio risks are high enough that the best probability of a 30 year retirement succeeding in the sense of no adjustments being necessary is about 80%.
I agree, there's no 0% failure plan and something like 80% is more realistic than 95%. Both in view of the likely gradual nature of 'failure' (it would usually leave room for adjustment over time, not like 20% chance of losing a dice game where you walk away penniless), and the fact that people would have to save so much more than the conventional metrics if they really wanted 95% in a more correct analysis IMO.

If I pull up FireCalc from
https://engaging-data.com/will-money-last-retire-early/
and put in very generic 4% SWR, 60/40, 0% expense standard SS mortality table, male 65 look at age 95, the success probability is 98%. If I put in what I believe is a minimal realistic very rough adjustment to expected return now v past geometric average with a 2% pa dummy expense it drops to 66%. But, it goes to 83% cutting the SWR fairly modestly to 3.6%. Would be less rosy if you figured a significantly longer LE than SS table, which most healthy/good health habits, well educated upper middle class 65 yr olds (typical on this forum) do. Then again SPIA's are priced attractively for such people. Anyway realistic (IMO) return estimates don't radically change the 'SWR' answer, from a fairly old starting age, if you accept that very high success probabilities are not realistic in the first place. But 5% probability from FireCalc can't be taken remotely literally IMO with now's valuations and yields. It's at most a relative measure, how relatively more risky this plan might be than that one, roughly.

I agree with your other post also that truly out to lunch optimism about withdrawal rates tends to be more serious in the FIRE/MMM contingent, that you could somehow have reasonable assurance of your money lasting from age 35 to death with no more work and withdrawal of 4% of initial plus inflation. OTOH MMM types are not necessarily retiring traditionally, they might plan things like moving from fast lane IT job (seems disproportionately) to slower lane physical trade, living in part off public benefits, strong belief they can switch back to the fast lane even decades hence (I'm skeptical but their risk), etc. It's a pretty different thing in some of those cases. Still, 4% in strict sense of that money alone lasting the rest of your life has an obvious major chance of failing over 50 or 60 yrs (though still a chance of more than succeeding. Low *expected* return is fundamental in view of a low global div yield and modest world GDP growth *expectation*, but a breakthrough that allows indefinite 6% world, or solar system, GDP growth would probably turn out very well for equity investors, since market doesn't expect that).
Last edited by JackoC on Thu Aug 04, 2022 11:31 am, edited 1 time in total.
petulant
Posts: 2225
Joined: Thu Sep 22, 2016 1:09 pm

Re: "Stocks increase SWR" assumes a risk-tolerant investor

Post by petulant »

JackoC wrote: Thu Aug 04, 2022 11:22 amLow *expected* return is fundamental in view of a low global div yield and modest world GDP growth *expectation*, but a breakthrough that allows indefinite 6% world, or solar system, GDP growth would probably turn out very well for equity investors, since market doesn't expect that).
Never bet against America the Solar Federation! - Warren Buffet, probably
nigel_ht
Posts: 4379
Joined: Tue Jan 01, 2019 10:14 am

Re: Watch out for hidden risk tolerance assumptions in SWR claims

Post by nigel_ht »

nisiprius wrote: Thu Aug 04, 2022 7:00 am (Background: SWR = "safe" (or "sustainable") "withdrawal rate," how much can a retiree withdraw annually from an investment portfolio which fluctuates in value, with little risk of running out of money. The "4% rule" is a common answer. The percentage is a percentage of the initial portfolio value. Once chosen, the withdrawals are adjusted for inflation, but not adjusted for investment performance).
Isn’t that baked in by assuming the worst historical investment performance?
The point I want to make is to be critical of claims that some portfolio change "increases the safe withdrawal rate," because whether or not it does may depend on a smuggled-in assumption about the investor's risk tolerance. All such assessments should be made on the basis of a targeted failure rate, and the choice of that target depends on the risk tolerance of the investor.
That’s a good point but I think many of the portfolio changes don’t increase equity allocations but increase diversification into size, geographic region, sector or asset type (*cough*gold*cough*).

My presumption is that international or SCV or cuz diversification generally reduces risk although often with reduced returns.

But the criteria for success is actually quite low given you need just enough gains to get from 25x expenses to 30x expenses to get from 3.9whatever to 4% and incrementally from 4% to 4.whatever%.

Given that SWR is dominated by local inflation rates anyway I don’t think AA moves the needle that much in increasing/decreasing total risk.

Not to mention that worse than historical outcomes are far less likely than worse than expected expenses.
Our crude exploration showed that an investor able to tolerate a 50% chance of running out of money is best served by a high stock allocation--100% stocks, in fact; but an investor only able to tolerate a 5% chance is almost indifferent to the allocation decision and is better served by 25/75 or 75/25 or 100/0.

(Or by 0/100. The success rates in Vanguard's simulation were 1% for 0/100 at a 7% withdrawal--1%! And only 86% at a 3.5% withdrawal).

Unfortunately, many SWR studies simply assume without any discussion that a 5% failure rate is appropriate.

Is it? That's a difficult question. Simply on the face of it, and as a planning assumption, if you were handed a pair of dice at the start of retirement, and told to roll the dice and accept that your chosen plan would fail if you rolled a twelve, would you accept that as a sound plan? That is not the only way to frame the decision, but it's relevant.
Well, most modern planning tools allow for a 100% historical success rate which is what I use…and 4% is better than 3.9something from a simplicity standpoint.

Given than SWR studies (published or informal blog posts) since Trinity have explored a wide range of SWR possibilities I think that the 4% rule is “good enough” in the same way my original 50/50 couch potato portfolio is “good enough”.
Last edited by nigel_ht on Thu Aug 04, 2022 11:41 am, edited 1 time in total.
nigel_ht
Posts: 4379
Joined: Tue Jan 01, 2019 10:14 am

Re: "Stocks increase SWR" assumes a risk-tolerant investor

Post by nigel_ht »

burritoLover wrote: Thu Aug 04, 2022 7:10 am The following chart gets thrown around a lot but shows much different results:

Image
I visit his site a lot but don’t interact with it…but I wonder if he would be amenable to updating his excellent SWR series with 2022 data…
nigel_ht
Posts: 4379
Joined: Tue Jan 01, 2019 10:14 am

Re: "Stocks increase SWR" assumes a risk-tolerant investor

Post by nigel_ht »

nisiprius wrote: Thu Aug 04, 2022 7:21 am It isn't "very different results." It actually illustrates my point. (There is a smuggled-in risk tolerance assumption in the choice of color coding). At lower withdrawal rates, everything is green over a broad range, all the way down to 25% stocks at and 40 years at least. The effect of "more stocks improves survival rates" is only true at the higher withdrawal rates, where nothing is very green.

The reason is pretty simple. If you choose a withdrawal rate that bonds can't sustain, then stocks will give you a shot at sustaining it--but not much certainty.
Well yes, but what that really means is that your retirement is underfunded given your desired expenses…by definition that’s riskier right?
nigel_ht
Posts: 4379
Joined: Tue Jan 01, 2019 10:14 am

Re: "Stocks increase SWR" assumes a risk-tolerant investor

Post by nigel_ht »

Tamalak wrote: Thu Aug 04, 2022 7:44 am It's a good post. I'm risk-tolerant so far - I didn't even flinch at the covid crash or this year's slump - but this is the accumulation phase. During accumulation phase bears are, in some cases, good, and even if they're bad all they do is shift your retirement schedule a bit.

I sense retirement will be more of a dependent, helpless feeling, since I'm not adding anything more to my portfolio. What will a covid lightning strike feel like then?
Nothing if you are on a 5 month vacation in bora bora without access to internet…of course you’d probably be trapped in bora bora unable to find a flight home…
User avatar
22twain
Posts: 3327
Joined: Thu May 10, 2012 5:42 pm

Re: Watch out for hidden risk tolerance assumptions in SWR claims

Post by 22twain »

Riprap wrote: Thu Aug 04, 2022 11:12 am How are the graphics posted to this thread?
Hit the quote button on any post with graphics and you'll see the code. I suggest the first post by BurritoLover with a graphic that you come to when you scroll upwards from this post. It's short and has one graphic, so you can easily see what's going on. Don't hit the "Submit" button on that page! Hit your browser's "back" button when you want to move on.
It's "IRMAA" (Income Related Monthly Adjustment Amount), not "IIRMA" or "IRRMA" or "IRMMA".
nigel_ht
Posts: 4379
Joined: Tue Jan 01, 2019 10:14 am

Re: "Stocks increase SWR" assumes a risk-tolerant investor

Post by nigel_ht »

vineviz wrote: Thu Aug 04, 2022 8:53 am
Probabilities are hard.

Many backward-looking SWR studies gloss over the fact that the worst REALIZED sustainable withdrawal rate was not actually the worst POSSIBLE rate that historic conditions could have produced.
Well, yes…because the worst Possible rate was 0% when the local exchanges was closed and your assets were seized.

These kinds of risks are not amenable to mitigation with traditional financial planning tools until you are in the 8 digit category of wealth…
In other words, the worst historical SWR was (based on my analysis) roughly equivalent to a 5% or 10% probability outcome using the kind of Monte Carlo simulations employed by nisiprius in the OP.
That may be true globally but not for every market.
In reality, planning for a truly 0% failure rate using forward-looking estimates would be insanely conservative whereas planning for a 0% failure rate using backward-looking outcomes is probably more aggressive than most people understand.
When planning for the military I was taught to consider most likely and most dangerous ECOAs (enemy courses of actions).

If you plan for most dangerous most of the time you wouldn’t be able to accomplish the assigned mission…or you had a “significantly disadvantageous exchange rate”.

Again, the kind of scenarios that drive the worst historical outcomes are not mitigable by methods commonly discussed in these forums.
JackoC
Posts: 3636
Joined: Sun Aug 12, 2018 11:14 am

Re: Watch out for hidden risk tolerance assumptions in SWR claims

Post by JackoC »

nigel_ht wrote: Thu Aug 04, 2022 11:37 am
Our crude exploration showed that an investor able to tolerate a 50% chance of running out of money is best served by a high stock allocation--100% stocks, in fact; but an investor only able to tolerate a 5% chance is almost indifferent to the allocation decision and is better served by 25/75 or 75/25 or 100/0.

(Or by 0/100. The success rates in Vanguard's simulation were 1% for 0/100 at a 7% withdrawal--1%! And only 86% at a 3.5% withdrawal).

Unfortunately, many SWR studies simply assume without any discussion that a 5% failure rate is appropriate.

Is it? That's a difficult question. Simply on the face of it, and as a planning assumption, if you were handed a pair of dice at the start of retirement, and told to roll the dice and accept that your chosen plan would fail if you rolled a twelve, would you accept that as a sound plan? That is not the only way to frame the decision, but it's relevant.
Well, most modern planning tools allow for a 100% historical success rate which is what I use…and 4% is better than 3.9something from a simplicity standpoint.

Given than SWR studies (published or informal blog posts) since Trinity have explored a wide range of SWR possibilities I think that the 4% rule is “good enough” in the same way my original 50/50 couch potato portfolio is “good enough”.
I applaud you for noting that what the historically base simulations give you is the *historical* probability. But I assume OP is talking about the forward looking probability. I don't believe there is a plan for 100% forward looking probability, but more importantly given where we start now with assets valuations and yields I believe the forward looking probability can be reasonably estimated much lower than say FireCalc's historical one. As in example I just gave, if I rig up FireCalc with a generic 98% historical success plan. I believe the actual probability of success of that plan is more like 2/3's. I'm not certain it's exactly 2/3's but no way would I plan thinking it was 98%, looking in the direction that matters, forward.

But this brings us to another (more optimistic) point about another embedded assumption in these simulations, particularly relevant for somebody at the midpoint of Graham's range of stock allocation, 50%. On the other 50% you really don't have to simulate anything. The simulations generally assume you buy long term nominal bonds (let's say it's 10 yrs) at year 0, sell them at year 1 when they are 9 yr bonds for another set of 10 yr notes and so on. That's easier data wise, and fundamentally that kind of rolling nominal portfolio can have a return advantage depending on term premium ('roll yield') as well as real or perceived correlation advantages with stocks. But, it opens you to big losses due to prolonged bursts of inflation that you would never earn back, and 'rebalancing into oblivion' if stocks just keep going down and don't 'bounce' the rest of your life. In reality you can buy a TIPS ladder with duration matched to your horizon (conventionally, a 15 yr avg duration for a 30 yr retirement). Nor do you actually have to 'rebalance'. You can have that 'Liability Matching Portfolio (with your Social Security benefit) separate from a 100% stock 'growth portfolio'. Not saying this is strictly superior, just that in that case your duration matched TIPS (which you could also trade for SPIA's later on) and Social Security are not subject to varying future market returns. The SS+LMP portfolio could give you near* certainty of an outcome you could live with, w/ stocks to try to get to the outcome you'd like.

*I'm not saying the chance of a general US govt default or just Social Security cuts is zero, there really is no zero looking forward, but it's not a direct function of market variables. And it's hard to discuss either in any detail here. I view both as highly unlikely, single digit % range even long run if I had to guess. I know some view a general US federal default as impossible but I don't agree, others view SS cuts as somewhat likely but I don't agree with them either.
nigel_ht
Posts: 4379
Joined: Tue Jan 01, 2019 10:14 am

Re: Watch out for hidden risk tolerance assumptions in SWR claims

Post by nigel_ht »

JackoC wrote: Thu Aug 04, 2022 12:16 pm
nigel_ht wrote: Thu Aug 04, 2022 11:37 am
Well, most modern planning tools allow for a 100% historical success rate which is what I use…and 4% is better than 3.9something from a simplicity standpoint.
I applaud you for noting that what the historically base simulations give you is the *historical* probability.
Yes…and that’s all SWR promises.
But I assume OP is talking about the forward looking probability.
I hope not, and I don’t think that’s correct given the OP.
I don't believe there is a plan for 100% forward looking probability,
I believe the 100% probability strategy is called “Dying young and leave a good looking corpse”.

I don’t recommend it though.
As in example I just gave, if I rig up FireCalc with a generic 98% historical success plan. I believe the actual probability of success of that plan is more like 2/3's. I'm not certain it's exactly 2/3's but no way would I plan thinking it was 98%, looking in the direction that matters, forward.
If I thought that using a SWR number with a 98% historical success probability had only a 66% real world chance of success I’d probably buy a lot more gold.

As I posted in my own thread I treat planning for retirement like I would bid a fixed price contract.

33% risk of failure would require me to hold so much management reserve that the numbers simply wouldn’t work out.
billaster
Posts: 408
Joined: Wed Apr 27, 2022 2:21 pm

Re: Watch out for hidden risk tolerance assumptions in SWR claims

Post by billaster »

It seems rather odd to insist on, say, a 95% success rate for money lasting 30 years when the odds of a 65-year-old even being alive in 30 years is only 10%.

So even with Peter Lynch's 7% withdrawal, which some call "unrealistic", your odds of being a broke 95-year-old are only somewhere around 5% to 10%.

It's sort of like insisting that airplanes be 100% safe when you are more likely to die driving to the airport.
H-Town
Posts: 4696
Joined: Sun Feb 26, 2017 2:08 pm

Re: Watch out for hidden risk tolerance assumptions in SWR claims

Post by H-Town »

nisiprius wrote: Thu Aug 04, 2022 7:00 am
The point I want to make is to be critical of claims that some portfolio change "increases the safe withdrawal rate," because whether or not it does may depend on a smuggled-in assumption about the investor's risk tolerance. All such assessments should be made on the basis of a targeted failure rate, and the choice of that target depends on the risk tolerance of the investor.

Our crude exploration showed that an investor able to tolerate a 50% chance of running out of money is best served by a high stock allocation--100% stocks, in fact; but an investor only able to tolerate a 5% chance is almost indifferent to the allocation decision and is better served by 25/75 or 75/25 or 100/0.

(Or by 0/100. The success rates in Vanguard's simulation were 1% for 0/100 at a 7% withdrawal--1%! And only 86% at a 3.5% withdrawal).

Unfortunately, many SWR studies simply assume without any discussion that a 5% failure rate is appropriate.

Is it? That's a difficult question. Simply on the face of it, and as a planning assumption, if you were handed a pair of dice at the start of retirement, and told to roll the dice and accept that your chosen plan would fail if you rolled a twelve, would you accept that as a sound plan? That is not the only way to frame the decision, but it's relevant.
I agree with the thesis.

As the opposite point, investors / retirees should not change their portfolio to make their money last. They should structure their spending and other factors to make their money last. Those items below can be great tools to help them:

1) Pay off mortgage and all debts so your annual spending consists of mostly discretionary cost vs. fixed cost. It will help you to decrease the spending during market downturn.
2) Have a holistic approach to insurance. Be sure to protect yourself in case of accident, illness, lawsuit, negligence, act of God, etc.
3) Be flexible with your spending and evaluate your withdrawal rate every year. All you need to do is to determine whether you have to tighten up your budget or to burn that dough. Don't tinker your AA every year.
Time is the ultimate currency.
nigel_ht
Posts: 4379
Joined: Tue Jan 01, 2019 10:14 am

Re: Watch out for hidden risk tolerance assumptions in SWR claims

Post by nigel_ht »

billaster wrote: Thu Aug 04, 2022 1:06 pm It seems rather odd to insist on, say, a 95% success rate for money lasting 30 years when the odds of a 65-year-old even being alive in 30 years is only 10%.

So even with Peter Lynch's 7% withdrawal, which some call "unrealistic", your odds of being a broke 95-year-old are only somewhere around 5% to 10%.

It's sort of like insisting that airplanes be 100% safe when you are more likely to die driving to the airport.
Yes, but who wants to be the unlucky 95 year old who outlived their money?
User avatar
vineviz
Posts: 12309
Joined: Tue May 15, 2018 1:55 pm
Location: Baltimore, MD

Re: Watch out for hidden risk tolerance assumptions in SWR claims

Post by vineviz »

billaster wrote: Thu Aug 04, 2022 1:06 pm It seems rather odd to insist on, say, a 95% success rate for money lasting 30 years when the odds of a 65-year-old even being alive in 30 years is only 10%.
It seems that most people would prefer to not outlive their wealth. Those people will rationally choose to plan for a retirement longer than their remaining life expectancy.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch
nigel_ht
Posts: 4379
Joined: Tue Jan 01, 2019 10:14 am

Re: Watch out for hidden risk tolerance assumptions in SWR claims

Post by nigel_ht »

H-Town wrote: Thu Aug 04, 2022 1:11 pm
I agree with the thesis.

As the opposite point, investors / retirees should not change their portfolio to make their money last. They should structure their spending and other factors to make their money last.
If diversification gives me another couple dozen basis points of SWR why wouldn’t I change my portfolio?

Just lowering volatility increases SWR…even with lower average returns.
User avatar
gmaynardkrebs
Posts: 2323
Joined: Sun Feb 10, 2008 11:48 am

Re: Watch out for hidden risk tolerance assumptions in SWR claims

Post by gmaynardkrebs »

nigel_ht wrote: Thu Aug 04, 2022 1:14 pm
billaster wrote: Thu Aug 04, 2022 1:06 pm It seems rather odd to insist on, say, a 95% success rate for money lasting 30 years when the odds of a 65-year-old even being alive in 30 years is only 10%.

So even with Peter Lynch's 7% withdrawal, which some call "unrealistic", your odds of being a broke 95-year-old are only somewhere around 5% to 10%.

It's sort of like insisting that airplanes be 100% safe when you are more likely to die driving to the airport.
Yes, but who wants to be the unlucky 95 year old who outlived their money?
I agree and under the best of circumstances, the healthcare costs at age 95 year old have got to be much, much higher than at 65 or 70, even if one is not in a nursing home. Seems to me one would want to plan for a "fat tail" of income needs after age 90, at the very latest. I wonder how many people do. I never really thought about myself, and I'm no youngster.
billaster
Posts: 408
Joined: Wed Apr 27, 2022 2:21 pm

Re: Watch out for hidden risk tolerance assumptions in SWR claims

Post by billaster »

vineviz wrote: Thu Aug 04, 2022 1:14 pm
billaster wrote: Thu Aug 04, 2022 1:06 pm It seems rather odd to insist on, say, a 95% success rate for money lasting 30 years when the odds of a 65-year-old even being alive in 30 years is only 10%.
It seems that most people would prefer to not outlive their wealth. Those people will rationally choose to plan for a retirement longer than their remaining life expectancy.
Life expectancy is 50% probability. These people are planning way beyond life expectancy to very small percentages of being alive combined with an even smaller percentage of being broke.

As we have seen with vaccinations, people can be highly irrational about probabilities.
User avatar
vineviz
Posts: 12309
Joined: Tue May 15, 2018 1:55 pm
Location: Baltimore, MD

Re: Watch out for hidden risk tolerance assumptions in SWR claims

Post by vineviz »

billaster wrote: Thu Aug 04, 2022 1:41 pm Life expectancy is 50% probability. These people are planning way beyond life expectancy to very small percentages of being alive combined with an even smaller percentage of being broke.
Planning on a retirement which exceeds life expectancy is precisely what people who understand probabilities SHOULD be doing.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch
nigel_ht
Posts: 4379
Joined: Tue Jan 01, 2019 10:14 am

Re: Watch out for hidden risk tolerance assumptions in SWR claims

Post by nigel_ht »

billaster wrote: Thu Aug 04, 2022 1:41 pm
vineviz wrote: Thu Aug 04, 2022 1:14 pm
billaster wrote: Thu Aug 04, 2022 1:06 pm It seems rather odd to insist on, say, a 95% success rate for money lasting 30 years when the odds of a 65-year-old even being alive in 30 years is only 10%.
It seems that most people would prefer to not outlive their wealth. Those people will rationally choose to plan for a retirement longer than their remaining life expectancy.
Life expectancy is 50% probability. These people are planning way beyond life expectancy to very small percentages of being alive combined with an even smaller percentage of being broke.
With the engineering risk management I am familiar with a very low probability but very high impact risk is considered a high risk that need a mitigation plan for.

And:

Image

8.7% that one person in a couple makes it to 100.

Perhaps you should reevaluate the probabilities…
billaster
Posts: 408
Joined: Wed Apr 27, 2022 2:21 pm

Re: Watch out for hidden risk tolerance assumptions in SWR claims

Post by billaster »

nigel_ht wrote: Thu Aug 04, 2022 2:09 pm
Image

8.7% that one person in a couple makes it to 100.

Perhaps you should reevaluate the probabilities…
Don't know your source for those calculations but they seem to be at least double the Social Security numbers.
User avatar
firebirdparts
Posts: 3578
Joined: Thu Jun 13, 2019 4:21 pm

Re: Watch out for hidden risk tolerance assumptions in SWR claims

Post by firebirdparts »

nisiprius wrote: Thu Aug 04, 2022 7:00 am I just want to illustrate a fairly obvious point. I don't believe it's controversial.
I agree with you here. I worry sometimes that people will find any reason to have a perverse discussion about SWR.

That is one of the more interesting outcomes of the study the way it was done. For the study to be useful, you have to to find a simplistic portfolio that wouldn't require clairvoyance, and really, he surpassed that.
A fool and your money are soon partners
User avatar
Nestegg_User
Posts: 1922
Joined: Wed Aug 05, 2009 1:26 pm

Re: Watch out for hidden risk tolerance assumptions in SWR claims

Post by Nestegg_User »

rocket354 wrote: Thu Aug 04, 2022 10:40 am My issue with "fail or not" analysis is that there is a huge middle ground that could be considered a failure that is not taken into account. If I'm left with $2.00 after 30 years with a given WR and AA...is that a success if I'm still alive?

One thing I noticed is that with higher equity allocations you get a slight bump in the failure rate in the mid-WR ranges (3.2-4.0%) but you also see a decrease, particularly in longer time frames (40+ years), in the "middling" outcomes--which I would cursorily define as having an inflation-adjusted balance less than the starting balance. These outcomes in many cases can be considered failures because one can be left stressfully withdrawing from a dwindling portfolio that you might actually outlive, and therefore can cause effective failures where one returns to work or decides to make dramatic changes to one's lifestyle to survive (the "eating cat food" outcome).

Ultimately, the decision is not, in my eyes, between:

90% success/10% failure
95% success/5% failure

It is potentially between:

85% success/5% middling/10% failure
75% success/20% middling/5% failure

Note that these numbers are made up for illustrative purposes. However, I think one can see that the decision in terms of AA and WR becomes murkier when the third outcome is introduced.
There's also a corollary of that, wherein the portfolio drops to a level that the retiree would "reasonably expect" they would run out of funds for their expected lifespan but the market delivered then above average returns and allowed for that case to "pass" versus the otherwise expected "fail".

When one reaches a similar condition in their own retirement, do they just assume "well, in this other simulation the markets gave them enough additional returns that it didn't fail" or do they do the more prudent action and find a "Plan B" adjustment that might actually work.
Post Reply