when to reset SWR basis

[musher]

I missed it if someone already made this point, but I think Grayhare's statement that they likely only have 8 yrs left to live is the most critical piece of information in the problem. When your maximum remaining lifespan gets short, prediction of the lifespan becomes more certain (in years), so one of the variables that the 4% withdrawal rate is intended to cover becomes more known.

I believe 6.25% at this point would not be too aggressive a withdrawal.

I'd think it through like this.

If I had a high degree of confidence my remaining lifespan was less than N, then I'd ask,

1. What's the largest portfolio decline I think is likely in the next N years.
2. Reducing my current portfolio by that amount, divide by N

That's my withdrawal amount.

For example,

Start with a portfolio of 1000 and an expected max lifespan of 8. Let's say, I expect a maximum potential portfolio decline (real $) of 50% in my remaining lifespan.
So Portfolio reduced by that decline is 1000*.5=500. Remaining amount divided by my lifespan is 500/8=62.5. That's a 6.25% withdrawal rate.

This doesn't work if you potentially have decades left to live, but when you're pretty sure you're in the last 10 years, it seems like a reasonable way to analyze risks in the remaining period left to you.

 

[flintnational]

I agree with you REWahoo. Let's say that I retire at 60 with a 30 year time horizon and use a 4% WR... my risk of failure is 5%. Then 3 years later... I reset/ratchet to 4% of my then higher portfolio.... at that point my risk of failure is less than 5% because my time horizon is only 27 years... so the failure rate might be something like 4.5% or something like that, but it would be less than 5%.

I'm guessing that one could probably "prove" this with Firecalc if one cared enough to bother to do so.

+1

Also, you are only resetting after the portfolio has increased. Accordingly, by definition, you have avoided some of the failure scenarios that start with portfolio losses in years 1, 2 or 3 (based on when you reset). I would think even without the shorter time horizon, you are at least statistically even with your starting survivability. Throw in the shorter time horizon and reset WRs should work.

 

[Chuckanut]

[-]
How is this any different from my friend Bill who retired a year later than I did with $1.1M? FIRECalc says he can withdraw $44,000 so why should I be limited to $40,000 plus inflation?

Maybe, so Bill will feel financially superior to you and buy you a nice lunch once a month.

Of course this works the other way also, right?

Suppose Bill retires a year later than you and due to a market downturn he 'only' has a measly $900,000. Now he can only withdraw $36,000 while you withdraw $40,000 + inflation. What happens then?

 

[REWahoo]

Suppose Bill retires a year later than you and due to a market downturn he 'only' has a measly $900,000. Now he can only withdraw $36,000 while you withdraw $40,000 + inflation. What will you do about that?

Buy him lunch once a month.

 

[flintnational]

Actually, Bill could pretend he retired the same year as you and have the same withdrawal. And, since he did not actually retire, his portfolio would be $40k greater than yours and he would have to buy you lunch once a month.

 

[REWahoo]

Actually, Bill could pretend he retired the same year as you and have the same withdrawal. And, since he did not actually retire, his portfolio would be $40k greater than yours and he would have to buy you lunch once a month.

Works for me!

 

[gcgang]

The 4% SWR seems like a test to see if you're able to retire.

Once retired, adjust as you feel comfortable.

I agree w OP that having more assets and less years to live several years into retirement allows for higher than original spending projections.

 

[RunningBum]

So let's say you retired in 1995 with $1M, so you can take out 25K the first year, increased by inflation after that.


The next 5 years the S&P did 37.6%, 23.1%, 33.4%, 28.6%, and 21.0%. Assuming you had a 60/40 AA, your portfolio did about half of that. Taking out 4% + inflation those years, your $1M turns into about $1.68M.


You really think that after 5 years like that, if you reset and take 4% of $1.68M your odds of success are unchanged?


I sure don't. I think someone who retired in early 2000 has as an increased risk as well.


Firecalc and the 4% "rule" are just guidelines. Common sense says that bull markets are eventually followed by bear markets, and the start of a bear market is a more dangerous time to retire than the start of a bull market. I wouldn't advise any one to retire today on a plan that requires 4%+inflation SWR for 30 or more years. I'd get a little buffer in case of a significant correction.

Is there anyone here who retired in early 2000, and has been spending a full 4% increased by inflation each year? How is that going, and how much were you sweating after 2002, and again after 2008? I'm betting at least some were making cutbacks or looking for some PT employment, which is not the same as success. You might've made it without adjustment, because the bear market wasn't

 

[flintnational]

I think you are correct. I may be double counting the avoidance of early failures and the shortened withdrawal period. Kites, in the article referenced earlier, probably has it about right. You can reset but only after a 50% portfolio increase from the starting value and then you only get a 10% increase.

 

[43210]

Our lawyer is innumerate and he said: "Take your stash and divide by the number of years you expect to live. Spend no more than that amount every year on average." So I am thinking we could live for 40 more years.

Your "innumerate lawyer" is over-simplistic, but is doing the first basic step to the calculation.

1/30~3.33% (adjust longevity as needed).

I really think that people coming up with ultra-low SWRs have forgotten about this basic first step.

 

[43210]

I missed it if someone already made this point, but I think Grayhare's statement that they likely only have 8 yrs left to live is the most critical piece of information in the problem. When your maximum remaining lifespan gets short, prediction of the lifespan becomes more certain (in years), so one of the variables that the 4% withdrawal rate is intended to cover becomes more known.

I think at older ages the longevity uncertainty is more, not less, if measured proportionately.

Looking at an Actuarial Life Table
https://www.ssa.gov/oact/STATS/table4c6.html
For males at birth, the middle 60% last 65 to 89 years.
For males at age 65, the middle 60% last 10 to 25 years.
For males at age 90, the middle 60% last 1 to 7 years.

It's true that the differences decrease with age
89-65=24
25-10=15
7-1=6

But the ratios increase with age
89/65~1.37
25/10=2.5
7/1=7

It is that increasing ratio that makes spending needs more uncertain as you get older.

 

[43210]

The 4% SWR seems like a test to see if you're able to retire.

Once retired, adjust as you feel comfortable.

I agree w OP that having more assets and less years to live several years into retirement allows for higher than original spending projections.

This. Just use 4% rule to make sure you have enough assets to retire in the first place. It would be very silly to literally use this rule once retired. It was never intended to be used that way.

 

[USGrant1962]

Your "innumerate lawyer" is over-simplistic, but is doing the first basic step to the calculation.

1/30~3.33% (adjust longevity as needed).

I really think that people coming up with ultra-low SWRs have forgotten about this basic first step.

Yes, I see this all the time - people advocating 2% as a SWR. 2% will last 50 years in TIPS pretty much no matter what. If 2% meets your needs or makes you feel good then go for it, but don't call it a Safe Withdrawal Rate (which in the research and literature means maximum safe withdrawal rate).

 

[pb4uski]

And recognize that if you can live on a 2% WR on a TIPS portfolio that it is likely that you worked a lot longer than you reasonably needed to.

 

[RunningBum]

Some people might have gotten to a 2% [-]SWR[/-] WR by working long enough to qualify for a hefty pension. It's kind of a reverse cliff. Work 29 years, you don't have enough. 30 years, your pension might cover most everything, and whatever else you had saved is more of a bonus.

It's also true that others, myself included, worked a bit longer than needed. It's not all on the assets side though, one could overestimate expenses.

 

[USGrant1962]

Some people might have gotten to a 2% SWR by working long enough to qualify for a hefty pension. It's kind of a reverse cliff. Work 29 years, you don't have enough. 30 years, your pension might cover most everything, and whatever else you had saved is more of a bonus.

It's also true that others, myself included, worked a bit longer than needed. It's not all on the assets side though, one could overestimate expenses.

But 2% is not their SWR, it is their WR!

 

[RunningBum]

But 2% is not their SWR, it is their WR!

Yes, sorry about that!

 

[USGrant1962]

Yes, sorry about that!

Sorry about the . I get aggravated by imprecise language - too many years an an engineer!

 

[RunningBum]

Sorry about the . I get aggravated by imprecise language - too many years an an engineer!

No, I get it, the misused "S" bothers me sometimes too, and yet I still added it. And I think you had just said something about this.

 

[SecondCor521]

Note that Greaney starts with a payout rate that he considers 100% safe. In his case, he can ratchet to another 100% safe payout without generating new potential failures.*

The "4% SWR" that gets thrown around here usually means a 95% success rate.
If you don't ratchet up with good early performance, you move your 95% success to 100% success.
If you do ratchet up, you give up your 100% success and go back to 95% success.

I'm not saying that one is better for everyone than the other. I'm saying that the decision to ratchet up is not "free", it involves trading away something of value.


* This assumes, of course, that there is such a thing as a 100% safe withdrawal rate.

This nicely encapsulates a disagreement I was having on another forum and now I understand why.

I agree with Greaney and plan to reset my payout periodically if my portfolio increases and as my lifespan shortens, using a WR that I consider 100% historically safe for my AA and predicted lifespan.

I think the person I was disagreeing with was mentally using a 95% safety factor, which leads one down the Russian Roulette analogy. In my case, I think that I'm playing Russian roulette but there are no bullets in the gun.

Personally I think that rate is close to 4%, which is the original studies of a 95% safe rate plus some income flexibility if needed plus some expense flexibility if needed plus SS plus some other backups.

One still does give up something, though. Even if you don't have a portfolio failure, the remaining expected portfolio value is lower in cases where one resets. This can either be viewed as the good version of "you can't take it with you" (as Greaney does), or leaving less of a legacy to one's heirs.

 

[musher]

I think at older ages the longevity uncertainty is more, not less, if measured proportionately.

That's true, I specified (years) in my original comment.

Longevity tables give you distributions for the population. Once we start talking about an individual, the tables can become less useful for prediction than the details of that individual's health. The older you are, the more true this becomes.

If you take Grayhare's self assessment of likely longevity at face value, it doesn't seem excessively risky to take an approach similar to what I suggested.

 

[ERD50]

... The reset doesn't change the odds. The odds are 95% because you are using historical market returns and those returns don't change simply because you made your first withdrawal from your retirement stash. Example:

I have $1M, run FIRECalc and it says I can withdraw $40,000. If, a year later my portfolio has grown, I have $1.1M and run FIRECalc, it says I can withdraw $44,000.

How is this any different from my friend Bill who retired a year later than I did with $1.1M? FIRECalc says he can withdraw $44,000 so why should I be limited to $40,000 plus inflation?

Agree with what you said and how you viewed it. But I think one can also view it from a different angle - if you look at the 5% failures, you can say that the people who retired in those years all had a 100% failure rate. And I'm pretty sure those failures would have been near (in hindsight) market peaks. So retiring near a market peak has a higher chance of failure than retiring near a trough. Though I think it's more useful to think of this in terms of "stress" on the portfolio, rather than binary pass-fail.

That's why I like to take a conservative 100% success rate, and say that even if my future is as bad as the worst of history, my portfolio will be stressed, but survive.

I agree with you REWahoo. Let's say that I retire at 60 with a 30 year time horizon and use a 4% WR... my risk of failure is 5%. Then 3 years later... I reset/ratchet to 4% of my then higher portfolio.... at that point my risk of failure is less than 5% because my time horizon is only 27 years... so the failure rate might be something like 4.5% or something like that, but it would be less than 5%.

I'm guessing that one could probably "prove" this with Firecalc if one cared enough to bother to do so.

You don't need to 'prove' it with FireCalc, it's self evident if you think about it. Easier to think of in terms of a 100% historical success WR (let's say that is 3.3% WR for the time frame) - if that rate passed in all scenarios in it's history, then it has to pass in... all scenarios in it's history. No matter when you start, 3.3% passed. So you can also ratchet up to 3.3% at any time, because that was also someone else's 'start time' and it worked for them. FireCalc doesn't know you from them. It's a little like the coin toss being 50%, regardless of previous tosses.

...

Suppose Bill retires a year later than you and due to a market downturn he 'only' has a measly $900,000. Now he can only withdraw $36,000 while you withdraw $40,000 + inflation. What happens then?

Bill could take the same amount. But FireCalc is always showing worst case in it's history, so it won't be able to 'tell' Bill that. That's what all those rising lines are. If FC reported the nominal success for everyone, all those lines would head near zero at the end date. It is conservative and gives the worst numbers.

-ERD50

 

[pb4uski]

[-]Isn't this true only if you do not take into account you are now using a survival target of less than the original 30 years?[/-]

^ After thinking about it, I disagree with myself. The reset doesn't change the odds. The odds are 95% because you are using historical market returns and those returns don't change simply because you made your first withdrawal from your retirement stash. Example:

I have $1M, run FIRECalc and it says I can withdraw $40,000. If, a year later my portfolio has grown, I have $1.1M and run FIRECalc, it says I can withdraw $44,000.

How is this any different from my friend Bill who retired a year later than I did with $1.1M? FIRECalc says he can withdraw $44,000 so why should I be limited to $40,000 plus inflation?

Au contraire... it does change the odds because the number of years is different but to your favor.

For example, in FIRECalc, input $40k spending, $1 million portfolio and 30 years and all rest defaults.... you get 94.9% success rate. Let's say that after 3 years despite withdrawals plus inflation, due to good market performance that the portfolio is now $1.2 million so you decide to ratchet and reset your withdrawals to $48k a year (still 4%). With a 30 year time horizon your success rate is still 94.9%... but you are 3 years older so the remaining time horizon is now only 27 years... and that change in time horizon increases your success rate to 96.7%.

Or if you want to reset to a 95% success rate but 27 years you can use the investigate tab to solve for the spending level commensurate with 95%... it is $50,824 or a 4.2% WR.

 

[Huston55]

As I understand it, the standard 4% SWR rule allows withdrawal of 4% of asset value during the first year, then permits annual adjustments in withdrawal amount to keep pace with inflation. This limitation of withdrawal increases to inflation provides protection against future downward moves in asset values.

Is there a point at which it is reasonable to reset the original basis of the 4%? With inflation low this decade at the same time stock prices have doubled and tripled, continuing to strictly limit annual withdrawal increases to inflation makes for tiny withdrawals relative to new asset values. If I started at 4% SWR in 2010, by 2018 my withdrawal, even with inflation adjustment, has become less than 2% of the assets. It seems some reset of my SWR basis is in order (no?), especially given that presumably I have 8 fewer years of life remaining.

GrayHare-

If you reset your WD, then you’re really using a VWR (Variable) instead of a SWR (which usually infers a set WD rate). No problem doing that but, we need to call it what it is. If that’s your preference, that also means you should be studying VWR methods to see which works best for you. There are many to choose from but, the two which are discussed most around here (and BH it seems) are: VPW & Guyton-Klinger (see Mr. Nut’s link @ post #31). My personal preference is G-K, which is what we use.

Looking forward to reading which method you choose & your rationale for the choice.

 

[RunningBum]

I guess we are in basic disagreement about the success rates after a strong bull market. I make a case for a possible bear market to follow, so reduced success rates. Some simply ignore that possibility and assume the odds are the same no matter when you start, because, you know, Firecalc. I'm willing to learn new things, but it's hard to learn when your point isn't being acknowledged. I'll take the hint and bow out. I have my plan that I'll stick with. I've seen nothing here to convince me to change anything.


btw, that 30 years starting point, then 3 years later means 27 left, doesn't really work out. If you make it 29 years, does that mean you will die the next year and that's all you should plan for?