Adjustment for using the year's high with the "4%" rule?

[bamsphd]

Most of the now classic "4% rule" style withdrawal studies I know about used year end values for their calculations. Year end is a rather arbitrary date, obviously chosen to make the calculations easier. Does anyone know of any studies that simulated the unlucky fool who happened to pick the year's high value as their initial withdrawal value?

I haven't felt this close to retirement since an October market high a few years ago. However, using today's close instead of the past year end close as my starting value probably means I need to adjust down my anticipated withdrawal rate. I'm just looking for any studies which can give me a ballpark estimate of how much of an adjustment I should make when [-]trying to convince myself that I like cat food[/-] doing my due diligence before going through one-way doors at work.

----
Disclaimer: The "4% rule" is simply a nickname for various systems of estimating the odds of various investment portfolios surviving for various periods using various withdrawal algorithms adjusted by various allowances for inflation based on various ways of specifying an initial portfolio value. All of the system are known to be at best rough approximations of the future, your mileage may vary, dried beans and rice is usually cheaper than quality cat food.

 

[rayvt]

The exact date probably isn't very important, as long as you use a 12-month period. However, the deadly occurrance is when your portfolio takes a big dip right after you retire and start withdrawals. If you arbitrarily move back to "recent high", you are purposely shifting yourself into that deadly scenario.

Look at the Guyton/Klinger papers. With the way they put "guiderails" around your withdrawals, they claim (arguably reasonably, IMHO) that you can use 5.5% to 5.5% as your SWR.

OTOH, there is already a waiting list for the job of "Walmart greeter" in my area, so you might want to think long and hard before you do something rash.

 

[DblDoc]

Are you asking - "Based on my portfolio at todays close the 4% SWR will allow me to retire now" then the answer is you should be fine. If your question is "I retired on 12/31/2010 and now want to bump up my withdrawal so it is 4% of my current, larger portfolio" then the answer is that is risky.

DD

 

[bamsphd]

Are you asking - "Based on my portfolio at todays close the 4% SWR will allow me to retire now" then the answer is you should be fine.

I'm saying that based on Friday's closed, for the first time since the crash, I'm feeling really close to being able to use a 4% withdrawal rate to satisfy our current spending rate. However, I'm not quite satisfied that 4% is safe enough, because:

• This sure feels like a bubble, and retiring near the top of a bubble is certainly not ideal.
• It would be EARLY retirement, especially for my younger wife, so 4% is probably too optimistic for a 50+ year withdrawal horizon.
• Medical inflation.
• The change in the family dynamics when instead of increasing spending in real terms every year, we lock in our current rate of spending.
• I need to update my calculations of how much we have actually been spending per year for the last few years. I'm scared I might not like the answer!
• There are some big ticket home maintenance items coming up fast.
• My wife does not want me to retire yet.

Last time I felt this close, I told my wife she could quit her well paid job to work for minimum wage at the local animal shelter. We found private insurance for her, and discovered that I would have to join the high-risk pool. Then the crash came, and dreams of retirement were postponed. Now I'm dreaming again.

I'm enjoying the dreaming. However, I'm nervous about what I will find when I go down the full check-list before jumping!

 

[Gone4Good]

However, I'm nervous about what I will find when I go down the full check-list before jumping!

Count me among those who consider 4% too high a withdrawal rate for many of the reasons you suggest.

To answer your specific question, I'm not aware of any calculators that fine tune the methodology to evaluate mid-year peak equity prices. While I suspect the outcome would be only modestly worse, it would be interesting to see how things would look with the assumption that you always retire at the worst possible time of any given year.

My bigger concern is that valuations for both equity and debt are in unusually high ranges, which make the results of any backcast model like FireCalc suspect. Taken together, I don't think there is any period in FireCalc's history where both bonds and stocks yielded as little (or had as high a multiple) as they do today. We're in uncharted territory, so caution is advised.

 

[accountingsucks]

Count me among those who consider 4% too high a withdrawal rate for many of the reasons you suggest.

To answer your specific question, I'm not aware of any calculators that fine tune the methodology to evaluate mid-year peak equity prices. While I suspect the outcome would be only modestly worse, it would be interesting to see how things would look with the assumption that you always retire at the worst possible time of any given year.

My bigger concern is that valuations for both equity and debt are in unusually high ranges, which make the results of any backcast model like FireCalc suspect. Taken together, I don't think there is any period in FireCalc's history where both bonds and stocks yielded as little (or had as high a multiple) as they do today. We're in uncharted territory, so caution is advised.

Count me in as someone who also thinks 4% is too high. I'm aiming for 3 to 3.3% (or 30 to 33X expenses). The big one you list are home repairs. I'd be alot more comfortable with a 4% SWR if I had just completed a whole bunch of big ticket home repairs (shingles, furnace, etc) + just bought and paid for a car versus 4% and facing those things. Perhaps at the very least aim for 4% + $x, x representing the anticipated house repairs.

 

[pb4uski]

Let's just remember that 4% is a general rule and is sustainable most of the time. If you took RE and shortly thereafter the markets headed south, could you rein in spending or do some part time work to reduce current withdrawals for a while until the markets recover? If you can say yes, then RE is realistic. If not, then there is risk.

 

[samclem]

Suggested changes in bold

Let's just remember that 4% is a general rule and [-]is[/-] has been sustainable most of the time over the last approx 140 years, in the US. If you took RE and shortly thereafter the markets headed south, could and would you rein in spending or do some part time work to reduce current withdrawals for a while until the markets recover? If you can say yes, then RE is realistic. If not, then there is more risk.

bmasphd, you are obviously giving this potential life change some serious study, which is great. I'm a little concerned (based on the "does it matter what the markets are doing on the day I retire) that you may be attributing more precision to the FIRECalc and other estimates than may be warranted.
Those who use the "4% of starting balance and then adjust for inflation annually" are extrapolating an unknowable slope off a present "bumpy" starting position. If you guys can live with the variation in annual spending under the "4% of each year's ending portfolio value" withdrawal model then you can be assured you'll never run out of money (though, if the annual withdrawal % + inflation exceeds your portfolio's growth rate, the true value of your annual withdrawals will slowly erode). Bob Clyatt's book provides a withdrawal method that smooths out the annual variability (with a floor at 95% of last year's withdrawal) which FIRECalc can model based on historic data.
I, too, think 4% is likely a bit too high for a retirement expected to last many decades. But, some folks plan to start out with a higher rate and reduce it when other income streams (e.g. SS, corporate pension) start arriving.

But, to answer your question: Nope, I don't know of any modeling for "4% based on beginning year market high" withdrawal amounts. I'm not sure it would tell you a lot. If you're going to use the "4% of starting value adjusted annually for inflation" method, you could instead use 12 month or 24 month average portfolio value rather than the value at your starting date to reduce the chances of being unlucky. I think you know that, based on the recent past market values, this would reduce your withdrawal check. Maybe that tells us something significant.

 

[Gone4Good]

Here's a thought on making an adjustment to FireCalc runs to reflect current valuations. The basic idea is to revalue your portfolio as if valuations were typical for the model's data set. That means a PE-10 for stocks of 16.4x compared to 23.4x today and a 10-year Interest rate of 4.7% compared to 3.3% today. To do that, simply reduce your equity balance by 30% and your fixed income balance by 11% and use the smaller portfolio for FireCalc runs.

So in the case of a $1MM portfolio split 60/40, the 'revalued' portfolio is reduced to $776K and a $40K annual withdrawal becomes a 5.2% draw assuming a portfolio valued right in the middle of FireCalc's data set.

 

[rescueme]

...do some part time work...

No way, Jose...

 

[REWahoo]

Here's a thought on making an adjustment to FireCalc runs to reflect current valuations. The basic idea is to revalue your portfolio as if valuations were typical for the model's data set. That means a PE-10 for stocks of 16.4x compared to 23.4x today and a 10-year Interest rate of 4.7% compared to 3.3% today. To do that, simply reduce your equity balance by 30% and your fixed income balance by 11% and use the smaller portfolio for FireCalc runs.

So in the case of a $1MM portfolio split 60/40, the 'revalued' portfolio is reduced to $776K and a $40K annual withdrawal becomes a 5.2% draw assuming a portfolio valued right in the middle of FireCalc's data set.

Rather than try to massage FIRECalc numbers, cut your expenses in half or keep working until your portfolio is 2x what FIRECalc says is "safe". Either way should work - and replace concerns about running out of money with alternative forms of misery.

 

[samclem]

So in the case of a $1MM portfolio split 60/40, the 'revalued' portfolio is reduced to $776K and a $40K annual withdrawal becomes a 5.2% draw assuming a portfolio valued right in the middle of FireCalc's data set.

Ouch. That math is bitter medicine. "I'd like to spin the wheel again, Pat."

 

[Andre1969]

I just did some quick Firecalc scenarios using a span of 59 years, which presumes that I retire right now, at 41, and die at 100. Here are the results I got...

3.3% SWR rate: 100% chance of success (none of the 82 59-yr cycles failed)
3.4% SWR rate: 98.8% (1 cycle failed)
3.6% SWR rate: 97.6% (2 failed)
3.7% SWR rate: 90.2% (8 failed)
3.8% SWR rate: 86.6% (11 failed)
3.9% SWR rate: 84.1% (13 failed)
4.0% SWR rate: 81.7% (15 failed)
4.5% SWR rate: 68.3% (26 failed)
4.9% SWR rate: 50.0% (41 failed)
5.0% SWR rate: 47.6% (43 failed)

And even at the 4.9% rate, where you have a 50/50 chance of making it, the worst cycle doesn't hit zero until about 18-20 years in. So it seems to me that if you're going to fail, it would be a slow-motion train wreck, with plenty of time to correct your course. Therefore, I wonder if it would be okay to start off with a higher SWR rate, but just be prepared to cut back if it doesn't look like it's going to work.

Also, for my calculations above, I didn't take social security into account. So, for most people, I imagine they'd have a better success rate than what I've posted above.

 

[rayvt]

Bob Clyatt's book provides a withdrawal method that smooths out the annual variability (with a floor at 95% of last year's withdrawal) which FIRECalc can model based on historic data.
I, too, think 4% is likely a bit too high for a retirement expected to last many decades.

I looked at Clyatt's "safe withdrawal" method/calculation, and I say UGH!!

It recomputes each year's withdrawal anew, taking 4.3% of that year's total assets, and putting a floor of 95% of the previous year's dollar amount. Your annual withdrawal "income" will vary wildly from year to year, but won't go below 95% of the previous year.

I see 2 major problems:
1) Rather than a steady progression of small "annual raises" like the standard "compute the initial WR amount once and then each subsequent year adjust that amount by inflation", you'll get a big raise when the market booms and then a series of 5% cuts when the market goes back down.
Wildly varying annual incomes are hard for most people to tolerate.

2) Sure, it uses a standard WR of 4.3%. But he includes in the calculation several things that are decidely non-typical. Like the assumed value that you could get by selling your boat, your motorcyles, second home, etc. And includes the "home equity surplus" you could realize by selling your house and downsizing to a cheaper house.
(Yeah, right----how easy is it to get a lot of money these days by selling those things? No buyers when the economy tanks, remember?)

Realistically, the example he shows on page 123 is not 4.3% withdrawal it's actually 4.6%.

As I see it, his WR method is simply a naive "take X% each year" computed on your portfolio value plus the phantom assumed value of your illiquid and questionable assets, with a floor of 95% of the previous year's amount.

To me, the Guyton/Klinger method makes more sense. Pick the initial rate (based on your actual portfolio value, not including any phantom asset values) and adjust by the annual inflation. This is the standard WD method. But then apply 20% "guiderails" to the IWR, and adjust your withdrawal amount up/down by 10% whenever your annual WR varies more than 20% of your initial rate.
For example, if your SWR is 4.5% your limits are 5.4% and 3.6%.

 

[Texas Proud]

IMO, if you are 'so close' to not knowing, then don't do it...


One of the major flaws in all these calculations is that they use 'market returns'... not many people earn market because we have to pay fees etc. AND we are not perfectly aligned with the market.... heck, I have not come close to a yearly market return my whole life... I am above some and below some others.... because I weight my investments differently than 'the market'... if you do the same.... you might or might not have a problem...

IOW, it is a rough estimate, not a science as to if you have enough... and you do not make a decision based on 4.00000%... like today you have a 3.9999% so you can not retire, but with a good day in the market you go to 4.00001% and you can...

 

[samclem]

I looked at Clyatt's "safe withdrawal" method/calculation, and I say UGH!!

Fine, everyone is different. Some people can tolerate significant changes in their annual withdrawals (because they can cut their expenses, or they can cover a larger part of their expenses with SS, a pension, an annuity, etc).

Here's why adjusting my annual withdrawals works for me:
1) Although I've structured my AA in a way I believe will keep up with inflation over time, I have no guarantee of that. My portfolio doesn't "know" what inflation has been. If I blindly take annual withdrawals based on inflation adjustments that are not linked to the performance of the portfolio (with or without "guiderails", guardrails, suspenders, gutters, etc), then I can easily drive my portfolio value to extinction.
2) If I withdraw more when the portfolio has done well and less when it has done poorly, I can help avoid selling shares when they are down in price. I think this is also the natural human reaction to market events, so I might as well model it.
3) I can accept annual changes in my withdrawal value to significantly improve the chances of preserving the value of my portfolio over time. Others may be in a different situation.

 

[Gone4Good]

One of the major flaws in all these calculations is that they use 'market returns'...

The other major flaw is that it assumes spending increases with inflation. I'm one who (after looking at historical data) thinks CPI is a very good measure of general inflation. But what CPI doesn't capture is the expanding basket of goods we buy. Twenty years ago, I didn't have a personal computer, or a cell phone, or internet access, or a whole host of other things that are pretty typical today.

I also think CPI understates our personal medical cost inflation. It may track the overall inflation rate exactly right, but our individual costs will increase just by getting older. Based on market quotes today my medical insurance premiums will increase 5% per year even if overall medical inflation is zero. If general medical inflation is closer to 8%, my personal medical inflation will very easily reach in to double digits. That doesn't even capture increases in out of pocket expenses I'll certainly endure as I age.

So the working assumptions imbedded in these models is that inflation is lower than what we'll likely experience and returns higher.

 

[mh]

Count me in as someone who also thinks 4% is too high. I'm aiming for 3 to 3.3% (or 30 to 33X expenses). The big one you list are home repairs. I'd be alot more comfortable with a 4% SWR if I had just completed a whole bunch of big ticket home repairs (shingles, furnace, etc) + just bought and paid for a car versus 4% and facing those things. Perhaps at the very least aim for 4% + $x, x representing the anticipated house repairs.

I always thought it was generally accepted that 4% should include
"depreciation costs"

 

[pb4uski]

I always thought it was generally accepted that 4% should include
"depreciation costs"

Agree. I include "normal" home repairs and auto depreciation in my budget and therefore in my annual expenses/withdrawals. Since I just renovated my retirement home I don''t need to worry about new roofs, furnaces, appliances, and such for a while. But the vehicles iwll ultimately need to be replaced.