Safe Withdrawl Rate -- 6%

[obryanjf]

I'll buy the argument (by Dory I believe) that a simulator based on constant purchasing power versus variable purchasing power is an apples and oranges comparision.

One of my issues with a variable purchasing power calc is I think you need to redefine failure. I coded his system (roughly) and some of the NPV's of the draws as they got adjusted down were 30%, 20% etc. of the inititial draw. Even though the portfolio never technically failed, it does not smell like success to me. I use 50% of initial draw as a failure, in my other simulators.

I like the idea of a couple of simple automatic adjustments (+ or -) after ER rather than a fixed 4%, but have yet to find the right rules. Some adjustments would be closer to mimic what I would want to do if things (portfolio) crashed.

BTW Guyton 6% Inflation Rule rule was eliminated in one of his follow on's in the paper, but I can't find it right now with kids hollowering and jumping around.

job

 

[sgeeeee]

You can use a 6% initial withdrawal rate and survive 30 years with 97% probability of success. You just can't inflation adjust the entire 6%.

Run Firecalc with a 60/40 portfolio, 30 year duration and a 3% withdrawal rate. I think I assumed short term govt securities. Then go down to the expense ratio box and bump it up to 3.18%.

This corresponds to someone who considers 1/2 their budget requirements to be essential and inflation adjusts them to make sure they have these things covered. They also withdraw 3% of their total portfolio value out each year for discretionary spending. If the portfolio skyrockets, they have a huge discretionary budget. If it tanks, they have less discretionary funds.

The first year in retirement, the withdrawal rate is 6%. After that, it is 3% inflation adjusted plus an amount that depends on portfolio performance.

 

[dory36]

I'm impressed! 

SGeeeeeeeeeeeeeee has come up with lots of things that the tool will do tnat I never thought of.

 

[2B]

I'll be impressed when I understand what happened.

If I put $1MM in as the "Starting Portfolio" and 3.18% as the "expenses," I get $29,600 as the starting withdrawl to be 97% "safe." That says that I have a SWR of $29,600 plus I get to take out 3% of my portfolio annually but the $ amount will vary with the size of the portfolio. In good years we travel to Italy and in bad years we don't. The mean inheritance for my children is cut to $1.3MM.

Carrying this forward, adding pensions or social security could then be put in and analyzed as part of the "safe" withdrawl portion that would be included in the inflation adjusted spending rate. The "expenses" could be adjusted to give more or less variable spending.

Is this right? If so, I'm very impressed with sgeeeee's post.

 

[Cb]

Have a look at this link in which SG describes the study in greater detail:

http://www.s152957355.onlinehome.us/cgi-bin/yabb2/YaBB.pl?num=1107790433

Cb

 

[cute fuzzy bunny]

Have a look at this link in which describes SG's study in greater detail:

http://www.s152957355.onlinehome.us/cgi-bin/yabb2/YaBB.pl?num=1107790433

Cb

Ick. You made me read something with a hosuc post in it

"Spidey Sense"

 

[ronin]

In the current study, Guyton eliminated the 6% inflation cap.  Also the withdrawal rule was modified with an additional stipulation that for a freeze to occur the portfolio had to have a negative yoy total return and the current WD rate has to be greater than the initial rate.  I made 2 spread sheets: 1 w/Guyton's rules and one w/Gummy's "sensible WD" rules and compared performances for various strings of returns.  I'd have to say overall the Guyton rules provide much more stability (less WD variance) at a slightly less cumulative draw, and ends up with more left over in the end.  It's interesting to jigger the initial WD rate down, and see what happens to the frequency of the triggger points too.  Personally, I like the principles involved in Guyton's rules.  Along with the flexiblilty we all would likely have to adjust, there's a lot to consider incorporating into my plan.

 

[sgeeeee]

I'm impressed! 

SGeeeeeeeeeeeeeee has come up with lots of things that the tool will do tnat I never thought of.

Dory is clearly the kind of guy I wanted to work with back when I had a job. He writes a fantastic program, makes it user friendly, and adds enough options to let you investigate just about anything you can dream of. Then he compliments you for filling in the boxes and pressing the run button.

Thanks for all the work on the program, Dory.

 

[sgeeeee]

You can also use FIRECalc to explore a spending model that includes regular reductions in spending level through retirement.

http://www.s152957355.onlinehome.us/cgi-bin/yabb2/YaBB.pl?num=1130395938/1#1

 

[2B]

Hey ronin,

What's the verdict? What would you specifically recommend?

 

[Jane_Doe]

I charged back to FIRECalc and ran the $40,000 withdrawl from $1,000,000 for 40 years with 5 yr treasuries.  That leaves an average residual of $8,266,000 (in today's dollars I think).  That's money I would like to have available to spend.
   

Hi all!

Interesting thread. Always fun to look at different scenarios and analyze different theories, even if I don't always follow it all!

Question for Soon2B: I tried to run your FireCalc figures of 40k SWR, 40 yrs on $1 mm port of 100% 5 yr. treasuries. I came up with 52% & 47% success rate (using ppi and cpi, resp.) Not trying to give you a hard time, but what other variables did you come up with to end up with an average of $8+ mm ending balance?

Also, have to agree with YearsToGo about belt and suspenders - our retirement will be entirely dependent upon our investments - no pension and small SS, so I figure very conservatively. For those of you who have pensions, high SS and other sources of income you may be more comfortable taking chances.

Another interesting note regarding ups and downs on spending - after 25 years of a flexible income (self-employed, husband had a small construction business), I can attest to the fact that it is just simple human nature to put off expenditures in lean years and spend extra, or buy those things that you put on hold, in the "fat" years. Easier than you think, really.

Printed out the whole Guyton article but haven't read it thru yet - probably will work wonders on putting me to sleep tonight!!

Have a great day!

Jane

 

[2B]

When I ran the case I used 60% stocks and not 100% bonds. The 60/40 split was based an earlier post by sgeee. I can see where you got the 100% 5 yr treasuries. I wasn't complete in my description.

You can use a 6% initial withdrawal rate and survive 30 years with 97% probability of success. You just can't inflation adjust the entire 6%.

Run Firecalc with a 60/40 portfolio, 30 year duration and a 3% withdrawal rate. I think I assumed short term govt securities. Then go down to the expense ratio box and bump it up to 3.18%.

 

[samclem]

To second rmark's observation--the complex rules look like an artifact of data mining. Any scheme to determine a withdrawal method/amount should result in general, simple guidelines since there's no precise future data to go on. Highly specific rules imply a degree of precision in our understanding of future conditions (asset class returns, inflation, etc) that is unwarranted (IMO).

Gummy's approach looks right to me.

samclem

 

[sgeeeee]

I think some people take the spending models too literal. I would be surprised if anyone actually uses a 4% withdrawal adjusted for inflation each year. It doesn't make sense. One year you have to re-roof the house, one year you have to trade cars, one year the air-conditioner goes out. . . Those kinds of occasional, unpredictable expenses cause fluctuations in spending that keep you from using a rigid budget -- even if you had nerves of steel during prolonged market declines.

But spending is the most powerful weapon in your retirement survival arsenal. Running simulations that look at different spending models can give you a good feel for how much you can gain by using different spending approaches. You can look at 1) inflation adjusted spending, 2) strict percentage of the portfolio spending, and 3) a mix of the two using FIRECalc. If you use the withdrawal change options, you can even look at declining or increasing spending models.

Real life will only approximate these models, but the results are still instructive.

 

[retire@40]

I think some people take the spending models too literal.  I would be surprised if anyone actually uses a 4% withdrawal adjusted for inflation each year.  It doesn't make sense.  One year you have to re-roof the house, one year you have to trade cars, one year the air-conditioner goes out. . .   Those kinds of occasional, unpredictable expenses cause fluctuations in spending that keep you from using a rigid budget -- even if you had nerves of steel during prolonged market declines.

That's why you build into your budget an allocation for those types of expenses so it can be part of your 4% SWR.

 

[unclemick]

True

My spending 1993-2005, has been all over the map when viewed SWR wise.

I view things like FireCalc as very useful guides and a valuable informational tool.

My bias - current yield - and change spending/and or what's in my portfolio as reguired.

heh heh heh heh

 

[cute fuzzy bunny]

Yep, as mentioned over the years, I put together an annual expense budget, then put together all the stuff I'll need to buy in 10 and 20 year periods. New cars, tires, furnace and air conditioners, kitchen appliances, furniture, etc.

Composite the annual "costs" by using the real expenses in conjunction with the amount I need to virtually "put aside" for the capital expenses.

If that is under 4% of my stash or my stash is 25x the annual composite cost, i'm good to go. Every once in a while when i'm bored I revise the budget spreadsheet and look at the net worth and see if i'm still better than 4%. So far after 5 years the number keeps improving.

So when the number is really a lot better and your investments have thrown off so much cash, you buy your wife some diamond earrings and a lexus and really put the pressure on the co-workers and neighbors husbands...

 

[SteveR]

If that is under 4% of my stash or my stash is 25x the annual composite cost, i'm good to go.  Every once in a while when i'm bored I revise the budget spreadsheet and look at the net worth and see if i'm still better than 4%.  So far after 5 years the number keeps improving.

So when the number is really a lot better and your investments have thrown off so much cash, you buy your wife some diamond earrings and a lexus and really put the pressure on the co-workers and neighbors husbands...

CFB,

Does your "better than 4%" include your DW's salary? If so, what would you say you were taking out of a real basis from your accounts? Just curious.

 

[cute fuzzy bunny]

Nope, I dont figure her salary in. Half of that goes to her 403b, about a third of the remainder funds our Roths, and the rest pretty much pays the monthly utility bills, gas and so forth. Remember she's only working a few days a week.

Now, should the **** hit the fan, we stop the 403b contributions, think about those roth contributions, and start pulling in the belt a little.

Real basis from our accounts? Wow...thats a hard one. I'm not really one of those quicken/adding up my receipts type. Let me do a little back of the envelope figuring in real time here...note that my financial situation has swung a little bit between being a single ER two years ago, married and having my wife on maternity leave for six months last year, to married to a part time working spouse and having a baby in the house this year.

Our funds pay out a weighted average of about 2, maybe 2.25-2.3% in dividends. What my wife brings in after taxes and other above mentioned deductions runs to about 1% of our total portfolio. At the end of the year I always have extra money laying around that I have to reinvest (or buy a car with). About 20-30% of the dividends paid out were left over last year.

So real consumption of the portfolio is dividend only, minus the bit I reinvest. We dont plan to sell any shares ever to fund our living expenses. I'll tweak the wifes deductions to fill any shortfalls or overages until the 'system' runs efficiently.

As my wife reduces her hours and eventually quits, our target retirement funds will have slowly increased their bond holdings, and therefore dividend payouts. Presumably I can tweak the output and she can tweak the hours to keep our funding flows 'efficient'.

Were she to quit tomorrow (I wish she would! HI HONEY! ) I'd put us right back to 50/50 wellington/wellesley and keep living off of the approximately 3.5% dividends thrown off, that'd keep us at about the same lifestyle indefinitely, with inflation protection.

 

[SteveR]

CFB...thanks for the response. It is helpful to see how you guys are doing it compared with all the "noise" here sometimes.

Thanks

 

[cute fuzzy bunny]

Shoot and I thought I was the big noise maker.

 

[LovesLife]

Composite the annual "costs" by using the real expenses in conjunction with the amount I need to virtually "put aside" for the capital expenses.

Would you "dumb this down" a bit? Not sure what this means Example? Thx.

 

[ronin]

Hey ronin,

What's the verdict?  What would you specifically recommend?

I'm still playing with the spreadsheets and data.  So far I've plugged in the CPI from '70-'05, and the returns of various AA using data from Fundadvice.com for the same period.  Depending on what you want to have happen (I used a 4% initial WD% for both, with a 25% bonus on the "gummy" excess profits), and what you'd be willing to put up with, the Guyton rules strategy is the clear winner.  The ups and downs are, comparatively, eliminated.  The 35 yr avg is larger, SD way smaller, cumulative amt greater and terminal value larger as well.  Of course, who knows what returns will look like for the next 35 yrs. but I think the general principles would still apply and results would likely by analogous.

 

[cute fuzzy bunny]

Would you "dumb this down" a bit? Not sure what this means Example? Thx.

I'm sorry LL...I made that a lot more complicated than it needed to be.

You can do it one of two ways.

Lets say you're going to need a car in 10 years. Looking at the price of cars and how much the price has changed in the last ten years, figure out roughly what you think the car you'll buy will cost in ten years.

Then take 1/10th of that and add it to your annual budget. But dont take it out of your portfolio.

If you're a real neatnik, figure out the amount you'd need to put aside that, with investment growth, net present value, tiered and indexing from year to year to accommodate the differences...whoops...theres that complication crap again and you really dont need it.

By factoring in a fat inflation factor and not figuring in the investment returns and all that other funny business, the results (I think) are pretty conservative).

Or I can save you a whole lotta work and effort. I listed these items for "decade spends", IE, what I'll need to buy in ten years:

2 Cars
Half a water heater
Half a furnace
Half an AC compressor
Housepaint
Half refrigerator
Half Dishwasher
Half range
Half Microwave
Half washer and dryer
Four sets of car tires
Two computers
Printer
Two tivos
Three televisions
Living room set
Bedroom set
Two gas grills
Replacement clothes
Replacement exercise equipment
Personal care appliances
Two Vacuum Cleaners
Two Steam Cleaners
Tools

and came up with ~ $50,000 worth of maintenance and replacement for long term stuff, or about $5000 a year in set-asides.

If you're a homeowner with a couple of cars and about the same needs, add $5000 to your annual expenses, and if thats more than 4% of your investment stash...uh oh...

This may be an eye opener for people thinking about doing $25k-30k bare bones possum living that arent figuring that stuff in.

If you screw up and forget something, or inflation runs wild, or the market takes a huge dump, you can defer some of these or buy a cheaper car or furniture, or go Goodwill shopping for stuff.

But its a decent broad rule of thumb.

We compared notes a couple of years ago and people who did the same analysis all were within walking distance of each other...I think ESRBob was the one who started that and I stole his list of stuff and added to it.

By the way, when I work up my monthly budget and my 'decade spends', I add 30% to the bottom line to account for taxes, unexpected inflation, and other 'funny business' that you cant predict. Again, if this results in your falling from '100% firecalc' to '50%', reconsider that plan. Cuz somethings going to happen thats unpredictable.

Now if you're as conservative with budget planning as I'm demonstrating above, perhaps you can move to a 6% withdrawal rate and come out the same. Since I inflated most peoples base expenses by 50$, why not the withdrawal rate?

Make sense now?

 

[LovesLife]

Thank you. Much better.