6% + SWR's

[obryanjf]

As a continuation of previous work, I have modeled 5 withdrawl models with the Ty Bernicke model and then have added extras, which are pension and SS.

The link for the previous discussion is here http://early-retirement.org/forums/index.php?topic=7648.15.

In rough and rounded terms for the 40 year sims, as I see it,  using Ty's model on the fixed yields 5% SWR's and on the Hybrid model yields 6.0+% SWR's.  When I add extras, all the models SWR's are increaded by about 0.5%

I have not included the extras graphs due to the each person having a different number and value of extras.  I have tried to be pretty generic in the values I have used but with extras that is not possible.


Failures for 40 year sim, 4% to 10% draws for all models.
This shows a specific simulation for all models, 40 years with starting withdrawls from 4% to 10%.  The fixed model starts to have failures at 5% and the other models start to have failures at 6%.  All models quickly fade away as the withdrawls are increased slightly as indicated by the sharp slope of all lines after the break point.



Family of Failures for all models from 10 to 50 year sims
This graph again is the point of all models over 10 to 50 years simes and what is plotted is the first point at which failures start to occur.  Decreasing withdrawl rates for longer simulation is what should be expected.


40 year Fixed model w/ Ty included and 5% draws starting
Now this shows all the withdraws (inflation adjusted)expressed as a % of the initial withdrawl.  So the starting value is 100% and thereafter upper or lower.  It decreases as the Ty model reduced draws over time, until 75 where the withdrawls stop decreasing with age.  Notice the tight distribution around the average. 
The dotted red line is the average withdrawl percentage over time and uses the right axis.  My interpretation is that I would get a bit nervous when that percentage is 10% to 15% to 20 % and increasing.


40 year Hybrid model w/ Ty included and 6% draws starting
This again shows all the withdraws (inflation adjusted) expressed as a % of the initial withdrawl.  So the starting value is 100% and thereafter upper or lower.  It generally decreases as the Ty model reduced draws over time, until 75 where the withdrawls stop decreasing with age.  With the Hybrid model, there is much more variability as the portfolio value determines part of the withdrawsl.  Also notice the average decreases initially until 75 and then increases to about the 100% mark.
The dotted red line again is the average withdrawl percentage over time and uses the right axis.  For the Hybrid model, the withdrawl % is lower than the fixed model.


DYODD

job

 

[Cb]

Thanks for sharing that Daddy-O!

I find that sort of analysis reassuring - my plan is to:

- diversify broadly (USTM, ScV, LcV, REITs, PCRIX, Euro, Asian, Int'l V, Emerging, TIPS, Total Bond), likely resulting in a fairly low SD

- use a variable withdrawal plan similar to what you call "hybrid"...taking perhaps 3% fixed plus another 1% floating

- but take only 4% at the outset in the event that future returns are appreciably lower than in the past.

- only after 5 years or so have passed would I consider a "portfolio reset", and then only if our SS projections seem more reliable than they do today.

Cb

 

[modlair]

If Social Security is in your future, please be sure you include it in your planning and firecalc runs and whatever. It has a profound effect on SWRs.

 

[cute fuzzy bunny]

Thats one thing I noticed, was that a small income stream, even one appearing in the future, does have an amazing, and immediate effect on SWR.

Putting in social security at 62 (17 years away) causes firecalc to give me an extra thousand bucks a year starting right now and still retain 95% safety.

Something I'm sure ESRBob wired into his book, although I still havent read it yet.

 

[Cb]

If Social Security is in your future, please be sure you include it in your planning and firecalc runs and whatever.  It has a profound effect on SWRs.

I do...but I first reduce the SSA's current projections by 25% (we're nearly 46 & 47) to be safe...I really don't think Congress has the stones to cut UMC retiree SS bennies by 25%, but I figure the resultant inflation and/or tax increases might have a similar effect on our investment returns.

Cb

 

[2B]

I do...but I first reduce the SSA's current projections by 25% (we're nearly 46 & 47) to be safe...I really don't think Congress has the stones to cut UMC retiree SS bennies by 25%, but I figure the resultant inflation and/or tax increases might have a similar effect on our investment returns.

Cb

I think the SS payment will be sacred. There will be lots of noise but no one will have the guts to change it because "old people" vote their wallet. Anyone who threatens social security will become toast. Even screwing with the "escalation factors" will be poison.

What I think IMHO will eventually happen is that it will all be treated as ordinary income. That will allow us "fat cats" with other income or pensions to pay our fair share.

 

[2B]

Something I'm sure ESRBob wired into his book, although I still havent read it yet.

He does. I doubt you would learn much by reading it because most of his points have been exhaustively hammered here. I do think it would be worth you reading it, if for nothing else, a little different perspective on some of the issues.

 

[ESRBob]

Thats one thing I noticed, was that a small income stream, even one appearing in the future, does have an amazing, and immediate effect on SWR.

Putting in social security at 62 (17 years away) causes firecalc to give me an extra thousand bucks a year starting right now and still retain 95% safety.

Something I'm sure ESRBob wired into his book, although I still havent read it yet.

CFB,
Hey at least get a copy from the library, since your home remodelling story is in it! (p 243 - you are "Ted and Cati")

Following up on 2B's point, the way I handled SS in the book's projections was as follows:

Given that for most ERs their SS is not going to be at the maximum level due to the shorter contribution period and

Given that everyone in 2005, at least, was worried that future benefits could come under the trimming knife, based around means testing (which would include income earned from assets if not the assets themselves) and that early retirees benefits might not keep up with wage inflation and

given that I spend a good chunk of the book telling people it is OK to take on a bit of part time work to make ends meet during ER,

and given that FIRECalc is there for us to do detailed calculations,

Therefore I basically chose a conservative approach and assumed that whatever Social Security you get at age 62+ can be used to offset your part- time work income. In other words, you can't do part time work forever, and you get to quit at around the time your Social Security kicks in. If SS is higher than the work income you have been earning, then use the difference for cushion.

I know this may seem overly conservative, and for those whose principal retirement income plan is to live off Social Security it may seem downright loony, but for most of us here, at least those who do some part time work, it seemed a reasonable and conservative shorthand.

 

[bosco]

Hey Daddy O

I'm sure you mentioned this in one of your previous posts, but what do you define as failure for a variable % withdrawal portfolio? If I take X% per year, so long as X<100, it seems like it can never fail. Do you arbitrarily call it failure when the pricipal goes below a certain value?

interesting stuff.

 

[justin]

Putting in social security at 62 (17 years away) causes firecalc to give me an extra thousand bucks a year starting right now and still retain 95% safety.

I'm glad you brought that up. I hadn't really thought about it in terms of increasing my initial annual withdrawal to that extent.

I ran the numbers for me. I can take an extra $2500 per year with a fixed withdrawal strategy, or $3600/yr with a fixed+1.5% variable hybrid withdrawal. That's a significant amount given my desired $53000 annual withdrawal!

These numbers are based on me and spouse each getting $10000/yr at age 65.

 

[cute fuzzy bunny]

Hey at least get a copy from the library, since your home remodelling story is in it! (p 243 - you are "Ted and Cati")

Yeah, I know...and I DONT WANNA BE TED!

I'm glad you brought that up. I hadn't really thought about it in terms of increasing my initial annual withdrawal to that extent.

See, every once in a while I actually say something that turns out to be correct. Its not that often, so you have to pay attention.

A little income stream showing up anytime in your strategy plan term really DOES produce a pretty profound change, even in todays withdrawals.

Now try the other hotbutton...do your age 62 social security amount in firecalc starting in the year you turn 62, then do the 'higher amount' at age 67. Figure it for max withdrawal at 95% survivability. Much nicer number to introduce the smaller income stream earlier than the larger income stream later.

Sort of like "compounding"...starting early has a heck of an impact...

 

[obryanjf]

I have called it a failure when the current draw was <50% of the initial inflation adjusted (IA) amount.  But more recently i have also put a floor on the cumulative lower adjustment.  Originally the floor was set at 80% of the initial IA withdrawl.  Then when you got a failure it was really a train wreck (true failure, run out of money, cat food, etc). Last couple of days I added an absolute IA number (40,000) rather than a % of the initial draw.

So there are 2 controls at work here.  use_lowwd enables a failure if the IA draw is < 50% of Initial draw.  It gets counted a failure, but the sim continues so the averages and stuff are right.  use_min sets a floor under which no draws will be allowed, i.e. the withdrawl will be propped up if the model calcs to something less than min.  So in effect if use_min is ON then use_lowwd is ineffectual becasuse when triggered, if it a real failure.  I can turn either or both on/off easy.  Usually now if turn off use_lowwd and use_min is set to something reasonable.

The only model (when no lower limits exist) which will never run out of money (true failure) is the pure variable.  All others could and do have true failures under the right conditions.

job

 

[cute fuzzy bunny]

I would note at this point that over the years we've had cases made (whether credible or not) that anything over 2% would be unsafe and unsustainable, that 6% is doable, and that anything too far straying from 4% is dangerous.

Perhaps we're just looking at the range of the very conservative to the very aggressive in planning.

 

[SteveR]

I would note at this point that over the years we've had cases made (whether credible or not) that anything over 2% would be unsafe and unsustainable, that 6% is doable, and that anything too far straying from 4% is dangerous.

Perhaps we're just looking at the range of the very conservative to the very aggressive in planning.

Maybe folks are trying too hard to turn an "estimated rate" into a "rule". Even the "experts" cannot agree on what the future will hold so why should we put such a sharp edge on an estimated number? It is a starting place to estimate possible outcomes based on current information and probabilities. It is a tool and not the Holy Grail. A tool is only as good the dude using it.

 

[bosco]

I have called it a failure when the current draw was <50% of the initial inflation adjusted (IA) amount. But more recently i have also put a floor on the cumulative lower adjustment. Originally the floor was set at 80% of the initial IA withdrawl. Then when you got a failure it was really a train wreck (true failure, run out of money, cat food, etc). Last couple of days I added an absolute IA number (40,000) rather than a % of the initial draw.

thanks. in the case of a variable withdrawal, failure is, of necessity, more arbitrary than the other systems. In my case, however, with a partially inflation-adjusted DB pension, I see it as the logical way to tap the nest egg, since the fluctuations in withdrawals will be somewhat dampened by the other income streams.

 

[Nords]

Maybe folks are trying too hard to turn an "estimated rate" into a "rule".

Phew, no kidding. If I'd seen this type of rule-making debate before I ER'd then I might've had an allergic reaction that would've sent me screaming back to the comfort of cubicles & rush-hour traffic.

Even the "experts" cannot agree on what the future will hold so why should we put such a sharp edge on an estimated number? It is a starting place to estimate possible outcomes based on current information and probabilities. It is a tool and not the Holy Grail. A tool is only as good the dude using it.

I think it's "paralysis by analysis".

If we ER "experts" are endlessly wrangling over the rules and unable to clearly articulate a simple set of financial guidelines, then we're too screwed up to have any chance of our portfolios surviving until SS & Medicare.

So if we ERs obviously don't know what we're doing, then any worker with a little concern over making the leap would have no incentive to do something about their own situation. It's a hopeless case, and the only solution is to keep working until you can clearly see the second or third decimal digit. Any less clarity would force a personal introspection that strays too far from the comfort of circular number-crunching.

"Paralysis by analysis" would be the one & only time where I could see the value of JG's "just do it" approach. However the messenger's credibility does not sell the advice. People have to kick themselves out of that loop-- or have some significant life event kick them out of it.

Maybe we should look to Jarhead's example. He managed to ER without even using FIRECalc!

 

[cute fuzzy bunny]

We're also seeing the effects of Analytics bumping into Drivers.

Amiables like me just have to go with the flow

 

[Cb]

I don't understanding the "stick-poking" at the folks modeling different variable withdrawal strategies. The fixed withdrawal rate studies that Trinity, Intercst, and FIREcalc were intended to model never felt quite right to me - i was certain that I'd trim my withdrawals a bit during market downturns, but had little idea how much this might benefit me.

When SG posted the results of his study of various fixed plus variable withdrawals I was frankly quite surprised at the benefit...in most cases SWR's were improved by 20+ percent. I'm grateful that SG, ESRBob, Guytron, and Daddy-O share these results. Keep them coming!

Cb
PS: FWIW, I intend to use a variable withdrawal strategy in retiurement, but will stay with a ~4% withdrawal in the early decades in the event those predicting lower returns are correct.

 

[HFWR]

Hosuc has argued ad-nauseum about the number being wrong...

The point of the Trinity study was to show what would work, based on historical market returns...

It's not the Bible, folks... :

 

[Rich_by_the_Bay]

A little income stream showing up anytime in your strategy plan term really DOES produce a pretty profound change, even in todays withdrawals...Sort of like "compounding"...starting early has a heck of an impact...

You are sooo right about that. Shouldn't be surprising really We preach the 4% and 25 x annual expenses rules so much that we forget the converse: any income stream we create is the arithmetic equivalent of having twenty five times that amount nest-egged (is that word?). 10 grand a year is like $250,000 in the market, at least from a cash flow angle.

Which also explains why pensions are so valuable. Any why they are disappearing.

Course with income streams, when they go away (e.g. part-time job), it's like losing that phantom "savings." I could now bring up the Immediate Annuity topic, but it's Friday and I am in a pretty decent mood .

 

[cute fuzzy bunny]

I think theres semiviolent agreement and honestly I dont think i've seen anyone say that 4% is immutable or "biblical". I've seen people get the business because they were jerks, because their idea just wasnt very good or because they didnt do a good job of explaining it.

I dont think anyone is saying that its 4%, only 4% and you're an idiot for looking at other ways.

I hear some people saying that its just not plausible to come up with a formula thats any more accurate than the horseshoes and handgrenade method.

I hear some people saying they think they can or should try.

Unless you've got a time machine or a really good crystal ball, you pick something that makes you feel good, a strategy that you think is going to implement your feel-good idea and you go with it. I just wouldnt expect a bunch of other people to jump in with you. Everyones got different buttons and different things happen when you push them.

Ten bucks says whatever chart or strategy you come up with, you're going to make a lot of changes and what you think's going to happen isnt.

The worst thing you can do, in my lame opinion, is to come up with a bunch of stuff that makes you think you can take a lot more out than seems likely, and you end up broke at 75. Second worst thing is being so afraid of "losing" your money that you live miserably and leave a ton of money at the end.

If you're so close that you need to push the edge to make a plan work, or you're so afraid of losing money...stay at work...its going to work out a lot better for you than the constant gnashing of teeth.

 

[justin]

When individuals like Daddy-O or CFB bring up different ways to look at withdrawal rates and things to make sure to add in to the FIREcalc calculator, I appreciate it. Since I'm very confident I won't have a pension or other lifetime streams of income (other than SS which may occur ~30 years after ER), my "nest egg" will have to do most of the heavy lifting by producing a stream of income.

If I can go with a hybrid withdrawal strategy instead of the fixed strategy, I need to save $250,000 (or 20%) less. That could be a few extra years of not having to work. And I get to spend more in the up years as the portfolio value (hopefully) rises. In the down years I can scrimp a little, or find a way to make a little money. To me, ~3.x% fixed versus ~4.x% hybrid is much more than semantics.

Of course many say "if you're that close to the edge, work a little more to pad your nest egg". Probably good advice, but I'd like at least some evidence as to what would have worked historically in order to know when I'm FI and not just some dude with a big 401k account balance.

There's a good chance I get to the point where I say, "according to FIREcalc, I'm FI", but I chicken out on pulling the cord and decide to work a little longer to pad the portfolio a little more. Maybe till I can live off the dividends and interest, who knows.

 

[SteveR]

I don't understanding the "stick-poking" at the folks modeling different variable withdrawal strategies. ..

Cb,

Maybe I am missing something but I don't see any stick poking going on here. I see an open and honest discussion of opinions on a topic that is one of the foundations for the existence of this board (FIRECalc). I know I have no malice towards anyone on this topic and I don't believe any of the other posters here do either. Frustration maybe...but not malice.

A model is a model...it is a representation of something else and is not itself the object it is intended to represent. SWR is a model of a set of probabilities based on assumptions. The possible variation in these assumptions is what makes one SWR different from another. In the end they are all opinions. We know all about opinions....."we all have one and mine is better than yours etc..... "

 

[obryanjf]

Wow, we got some good discussion going now.. that great. I agree with most of what everybodies said actually.

I'm interested in understanding and kicking the tires of many differing options/models and picking the one that makes sense for me. Other people may have different needs, income streams and a different model may fit their lifestyle and temperment better. So be it.

I'm a DIY person and would not feel confortable unless I had a plan and model of the likely outcome. Some jump into ER or R without knowing or understanding the likely outcome; some are lucky, some have an intuitive financial sense and some we hear about on Frontline.

I also realize that if so and so model yields a 6.23 or whatever SWR, future returns will make it a different number, but likely somewhere in the ballpark. Someone else on another thread mentioned that the NUMEBR is the maximum SWR. In my case, playing with the models is helping me to reevaluate that 'maybe' i'll reduce my number from 1.25MM @ 4% to 1.00MM @ 5% even though my models say I could draw 7 or 8% (with Ty@55, Pension@65, SS@67, DWSS@72 and partime to stay busy). cushion built in. That equates to about 3 to 5 years earlier than just taking the conventional wisdom.

Just like I do not or would not trust my future soley to the word/work of ; financial planners, the trinity guys, Guyton, T Rowe Price calculator and as much as he seems like a great guy, dory and FIRECALC; I have the ability and means to plan for myself and will.

job

 

[REWahoo]

...I have the ability and means to plan for myself and will.

job

Yep. There are those like JG who swear the "SW" in SWR stands for Smith & Wesson.