FIREcalc criticized in blog

[gindie]

I apologize if I am bringing up old battles here.

I subscribe to a publication called Retirement Weekly, put out by MarketWatch.com (formerly CBS MarketWatch).  This forum was mentioned in response to a call by the editor for readers' favorite retirement sites.  A link to the forum, along with a special mention of FIREcalc, was published in the July 15 edition.

Another mentioned site was passionsaving.com.  I went to this site and read the blog at http://www.passionsaving.com/Financial-Freedom-Blog.html

That author claims to know our host and has nothing but great things to say about this bulletin board.  However, he criticizes FIREcalc quite a bit, claiming it is based on faulty SWR assumptions.

I am interested in reactions to this criticism as I always thought FIREcalc was a pretty good tool.  Again, if this is a rehash of old stuff on the board, then I apologize.

 

[wabmester]

The author is an interesting character.   Some would say he's waging his own little personal Jihad against FIREcalc.   Others would say he's simply psychotic.  In any case, he's the only person I'm aware of to have been banned from this board.   In fact, he's been banned from nearly every board to which he posted.

You can read all about him (sometimes in his own very verbose words) here:

http://www.retireearlyhomepage.com/cgi-bin/yabb/YaBB.pl?board=HOCO

You can boil down his rant against FIREcalc pretty simply.   He thinks the market is overvalued and that you should reduce your withdrawl rate during times when he thinks it's overvalued.   What he doesn't seem to get about FIREcalc is that it includes all the historical US stock data returns, including severe downturns like the Great Depression, and then it considers the *worst case* sequence and tells you if your portfolio would have survived that sequence at a given withdrawl rate.     That's all it does, and it does that fine.

 

[Nords]

The author is an interesting character.   Some would say he's waging his own little personal Jihad against FIREcalc.   Others would say he's simply psychotic.

I think he's both. The fact that he objects to FIRECalc is an endorsement.

I can't believe he duped MarketWatch into promoting his crap. The fact that he duped MarketWatch says a lot about their due diligence and the quality of their financial advice...

 

[unclemick]

Here's my take - I run FIREcalc all the time to compare various options/scenario's - I JUST DON'T USE IT FOR RETIREMENT. Going into my 12th year of ER with a no. 2 pencil and a single sheet of engineering graph paper - the back of the envelope works just fine.

My central heartburn with all calculators - historical/Monte Carlo and the old return tables based on interest type calculations - is their disrespect of my sweetie by the mailbox - the Norwegian Widow waiting for her dividend checks.

In the end - my portfolio is like the mythical - Sampo - the magic mill - it's what it cranks out in div.'s/interest that counts.

Ok, ok - I'm willing to concede - someday I 'may' get old - like 75 or 80 something - and be willing to sell parts of equipment off the old mill.

But not yet.

Heh, heh, heh, heh.

 

[bosco]

In fact, he's been banned from nearly every board to which he posted.

You can read all about him (sometimes in his own very verbose words) here:

http://www.retireearlyhomepage.com/cgi-bin/yabb/YaBB.pl?board=HOCO

To elaborate further, the other board that is accessible from the Retire Early Homepage is the only board that still allows him to post. I don't go there much anymore because he hijacks nearly every discussion with his obsessive, verbose, and sometimes nonsensical posts.

Bosco

 

[Yipee-Ki-O]

Hey unclemick, I'm hoping you let your Norwgian widow sweetie flirt a bit with other fellas . . . 46% of my net worth is in individual dividend paying stocks so I plan on hanging out by that magic mailbox sometime soon myself!

 

[maddythebeagle]

ok. I was wondering why that "other board" is so empty. I took a little time to look at this guy's website. I am not really impressed and he looks to be another joker just trying to sell a book. :

 

[lazyday]

What he doesn't seem to get about FIREcalc is that it includes all the historical US stock data returns, including severe downturns like the Great Depression, and then it considers the *worst case* sequence and tells you if your portfolio would have survived that sequence at a given withdrawl rate. That's all it does, and it does that fine.

Of course that means that if the market does worse in the future than it ever did, one might run out of money with a w/r that shows safe in Firecalc.
For example, some might say that around 2000, the market was in unchartered terrirory, more overvalued than ever before, and that a withdraw rate that always had been safe might not be safe for someone retiring in 2000.
H* aside, I don't think that argument is crazy.
One might even be concerned that from today's market, a retiree wouldn't be safe with the lowest w/d rate that has historically been safe. That is more of a stretch, but might be a valid concern.
There is a survivorship bias in looking at an existing market that has been very successful so far.

There's other issues too, like using inflation rate vs wage growth rate causing the risk of falling behind on standard of living. And who knows how long future human lifespans might be.
My "solution" is to take Firecalc results with a grain of salt, and try to use a slightly lower rate. There's so many unknowns anyway, I don't think we should be taking the Firecalc results as precise numbers to be followed exactly. If one can afford a conservative w/d rate, over time the pot can grow, and standard of living could improve.

On the other hand, being too conservative could cause one to work too long, and waste time. The future might be much better than the past. In fact, probabilistically speaking, I think it will. I just don't count on it.

Anyway, I think this board is a much better place since the ban. I didn't even really start posting here much until then.

 

[wabmester]

Of course that means that if the market does worse in the future than it ever did, one might run out of money with a w/r that shows safe in Firecalc.

That's true, of course. All you can do is take FIREcalc at face value. It shows how you'd do in a historically worst-case sequence. Personally, I think it's *very* likely that we'll see worse long-term sequences in the future, but I couldn't tell you how much worse or what the probability is that you'll see one of those worse-than-the-worse sequences during your retirement. Nobody can. And, AFAIK, nobody can give you a better estimate of a SWR either. Although, 0% is probably guaranteed to be safe.

 

[Cb]

Of course that means that if the market does worse in the future than it ever did, one might run out of money with a w/r that shows safe in Firecalc.

There's other issues too, like using inflation rate vs wage growth rate causing the risk of falling behind on standard of living.  And who knows how long future human lifespans might be.

I think those are reasonable concerns - my plan is to:

- leave a fair amount of fat in my RE budget, around 25-30% higher than my "go back to work" threshold

- use a 55 year withdrawal period (~5% probability of one of us making it that far)

- retire when FIREcalc indicates a 95% success rate (the "combined rate" is then >99%)

- diversify much more broadly than FIREcalc's S&P and traditional FI options

- use a semi-variable 4% withdrawal plan like Salaryguru described...approximately 2.5% fixed (adjusted for inflation) and an additional 1.5% of portfolio

...and if it looks like we'll still come up short I'll get a low-stress job to fill in the shortfall

Cb

 

[lazyday]

And, AFAIK, nobody can give you a better estimate of a SWR either.

Agreed. I suppose someday, someone might come up with a better method. I wouldn't drive myself nuts trying though.

 

[unclemick]

Someone did - Da Norwegian widow.

Of course if you are a slice and dicer - lean more towards De Gaul.

If you like Ben Graham - pssst Wellesley.

Bogle used to quote Lord Keynes on history - something to the effect that history is wonderful - provided you understand the reasons why it was what it was.

Today the central difficulty is that market during the 'bubble' broke the historical trend channel to the upside. Sooooo - a good thing is a bad thing - right

As that great guru Yogi B. pointed out -'the future ain't what it used to be.'

As of 7/22/05:

Vanguard Balanced Index - 2.63% current yield
Vanguard Wellesley - 3.75% current yield.

Except for MY(you can't have it) no. 2 pencil - FIREcalc and a handgrenade are as good as it gets - plan wise.

 

[Cut-Throat]

I think FireCalc is the best type of planning tool that we have. With that said almost everything that is forecasted is based on past results. And even if the future is worse than the past, there are plenty of other 'safety valves' that can be utilized for a retiree.

For Example: Let's say that the U.S. does go into another Depression. How many retirees would continue spending their Full SWR? I know I could cut my budget in half, if I had to. Travel and Entertainment would go. Eating out would also, etc. etc.
But Firecalc plods along at the same spending pace.

I have suggested to Dory that a 'safety Valve' such as this could be built into FireCalc. Like Cut Spending X% after a Y% Market Downyear. Increase X% Spending after a Y% Market Upyear. I'll bet the SWR could be increased substaintially!

Also, there are Reverse Mortgages and other sources of income beside work.

Even H*cus admitted on this forum, that if you 'survive' your first 10 years of retirement without a major meltdown of your portfoilo, that there is not much chance of going broke with the FireCalc Recommended SWR.

Remember the worst case scenario of FireCalc is based on retiring immediately before the Great Depression. The first years of Retirement are the riskiest. IOW Run FireCalc with a 45 year Withdrawal period for a 95% success rate and you will have a Negative Number at the end of a 1929 Retirement. Look at the ending number for the year 1919 Retirmement. The number is probably huge! - You still experienced the stock market crash and the Great Depression, but it happended 10 years after you retired, instead of the same year you retired. - This is a HUGE difference, and one that most everyone ignores when running FireCalc.

 

[RockMiner]

My "solution" is to take Firecalc results with a grain of salt, and try to use a slightly lower rate. There's so many unknowns anyway, I don't think we should be taking the Firecalc results as precise numbers to be followed exactly. If one can afford a conservative w/d rate, over time the pot can grow, and standard of living could improve.

This is my thinking as well.  I plan to shoot for a 3.5% SWR in the first few years...we are coming through a period of extreme market behavior and the world economy is changing faster than I can read about it.  I see nothing sacred about 4%, there are way too many variables.  Plus I'm chicken   

 

[ataloss2]

may be beating a dead horse here but this seems to be a matter of what a calculator can tell you. That is what would have worked. Predicting the future isn't really something that you expect.

 

[dbr]

Maybe an interesting addition would be to calculate the joint probability of running out of money and still being alive. For example a sixty year old contemplating retirement could multiply the failure rate from FIRECALC for a 30 year duration (say 5%) by the chance of being alive at age 90 (say 10%) to get a probability of disaster (being alive and having no money) of only 0.5%. In actuality some sort of accumulated percentage needs to be calculated as once one runs out of money or once one dies at any age, one does not then recover. In other words, the conditional probability porblem needs to be correctly formulated. I am simply suggesting that running high rates of financial survival to time periods with low rates of life survival is kind of a strange approach to retirement.

If one is going to do statistics on the finances one should also do statistics on the longevity, it seems to me.

 

[wabmester]

Remember the worst case scenario of FireCalc is based on retiring immediately before the Great Depression. The first years of Retirement are the riskiest.

I think you're reading way too much into the historical data. The results you get from FIREcalc are due to very specific sequences of peaks and valleys. We were very lucky that the two big valleys (the Great Depression and stagflation of the 70's) were soon followed by enormous run-ups. I think you should keep in mind three things when thinking about the future:

1) Your retirement could be screwed by a longer or deeper valley (think Japan, for example).

2) Your retirement could be screwed by shorter bull runs (the 20-year run from the 80's that got many of us to ER is pretty unprecedented).

3) Your retirement could be screwed by a long-term slow down in the economy, as would be suggested by various factors such as our aging boomer population, shift to lower-paid service jobs, etc.

Basically, you shouldn't draw any conclusions from the past unless you are certain that they extrapolate to the future.

 

[Cut-Throat]

I think you're reading way too much into the historical data.   The results you get from FIREcalc are due to very specific sequences of peaks and valleys.    We were very lucky that the two big valleys (the Great Depression and stagflation of the 70's) were soon followed by enormous run-ups.   I think you should keep in mind three things when thinking about the future:

1) Your retirement could be screwed by a longer or deeper valley (think Japan, for example).

2) Your retirement could be screwed by shorter bull runs (the 20-year run from the 80's that got many of us to ER is pretty unprecedented).

3) Your retirement could be screwed by a long-term slow down in the economy, as would be suggested by various factors such as our aging boomer population, shift to lower-paid service jobs, etc.

Basically, you shouldn't draw any conclusions from the past unless you are certain that they extrapolate to the future.

Any form of planning is subject to variablilty. You still have to plan.

 

[wabmester]

I am simply suggesting that running high rates of financial survival to time periods with low rates of life survival is kind of a strange approach to retirement.

Not really.   Retiring early requires a survivalist attitude.   We're sort of cutting off our supply of oxygen and seeing how long we can hold our breath.   Assume you will live to be 120 (maximum human lifespan).   Assume that we have a repeat of the Great Depression.   Does your portfolio still cover your expenses over your entire lifespan?   Great, then *maybe* you can get away with retiring early.

 

[unclemick]

Yep

You can't always predict - the roll of dice history will deal.

In 1993 - I had 'NO IDEA' we would get one of the best historial decades of return.

Went merrily along as a really cheap bastard for 10/12 yrs - thoroughly enjoying our ER and ended up with a current - er ah perhaps excessive margin of safety - history/passage of time and all that.

Lightening up on budget control as time passes - BUT my heart of hearts enjoys - getting my jollies off on extreme frugal.

Things change/time passes. Have a plan - and be prepared to adjust/replan as things change.

Remember Bear Bryant's Alabama linebackers - 'agile, mobile and hostile.'

 

[intercst]

Maybe an interesting addition would be to calculate the joint probability of running out of money and still being alive.  For example a sixty year old contemplating retirement could multiply the failure rate from FIRECALC for a 30 year duration (say 5%) by the chance of being alive at age 90 (say 10%) to get a probability of disaster (being alive and having no money) of only 0.5%.  In actuality some sort of accumulated percentage needs to be calculated as once one runs out of money or once one dies at any age, one does not then recover.  In other words, the conditional probability porblem needs to be correctly formulated.  I am simply suggesting that running high rates of financial survival to time periods with low rates of life survival is kind of a strange approach to retirement.

If one is going to do statistics on the finances one should also do statistics on the longevity, it seems to me.

It's been done.

http://www.retireearlyhomepage.com/swrlife.html

intercst

 

[Cb]

I have suggested to Dory that a 'safety Valve' such as this could be built into FireCalc. Like Cut Spending X% after a Y% Market Downyear. Increase X% Spending after a Y% Market Upyear. I'll bet the SWR could be increased substaintially!

It's been done (thanks to SG):

http://www.retireearlyhomepage.com/cgi-bin/yabb/YaBB.pl?board=SWR1;action=display;num=1107790433

 

[dory36]

I have suggested to Dory that a 'safety Valve' such as this could be built into FireCalc. Like Cut Spending X% after a Y% Market Downyear. Increase X% Spending after a Y% Market Upyear. I'll bet the SWR could be increased substaintially!

I am wondering about a conceptual discrepancy when Firecalc is combined with a longevity calculator.

As several have mentioned, Firecalc is a worst case planning tool, not an estimating tool. For those new to these discussions, let me use an analogy:

If someone were to claim an ability to predict the temperature in St Louis for August 13, 2007, he would universally be recognized as a charlatan. But if someone were to say that, based on recorded history, you could be safe leaving your heavy winter clothing at home that day -- everyone would see that as a blinding flash of the obvious.

Could there be a new ice age? Sure. But not too likely.

THAT kind of prediction is what Firecalc and Intercst's spreadsheets offer.

Now, turn to longevity. Now we are tring to predict a lifespan. Just as problematic -- everyone is different, medical advances, etc.  Besides, there is good reason to believe that the E-Rs will outlive their otherwise identical employed counterparts, for lots of reasons which can be explored in a separate thread.

So to stay with the model Firecalc uses, we'd have to incorporate a"worst case"  longevity calculator that says something along the lines of "there's negligible chances of you living beyond age 110."

I think that most of the longevity factors that are used in most models will affect the mean lifespan, but because of the high variance, will be pretty useless for predicting individual cases.


Besides... You have to understand that the bulk of Firecalc was built while I was still working, and thus had plenty of time to do such things.

All my spare time has vanished since I retired.  To make major changes now would require that I go back to work, so I'd have more free time!

 

[MJ]

Besides... You have to understand that the bulk of Firecalc was built while I was still working, and thus had plenty of time to do such things.
All my spare time has vanished since I retired. To make major changes now would require that I go back to work, so I'd have more free time!

Hey dory, maybe retiring wasn't such a good idea, if it eliminated all of your spare time.

 

[unclemick]

Hmmm

Take another look at Dory/grandchild posted on the other thread - and then decide if you really want to ask that question.

Heh, heh, heh

Perhaps FIREcalc can wait.