Pfau on the 4% SWR question - yet again

[REWahoo]

Maybe I'm missing something, but it appears to me Wade Pfau is putting out conflicting information about the safety of the 4% WR.

February 2014:Forget the 4% Withdrawal Rule

“The probability that a 4% withdrawal rate will work in the future is much lower,” he says. His new safe starting point: a 3% drawdown.

Now this in July 2014:Retirement: 2 different views on the 4% rule

View #1: A histogram for historical outcomes

In this view, as 4% is the worst-case scenario from history, it can be reasonably viewed as a safe withdrawal rate for today's retirees.

View #2: Adding retirement date market valuations to the picture

...a best guess about the sustainable withdrawal rate for someone retiring today is 4.2%....

Maybe I'm missing his point but in addition to being substantially different from what he said in February, the two views seem to be entirely supportive of each other. And his closing line only confuses me further:

Which view is most relevant is the issue plaguing today's retirees.

Makes me question his credibility.

 

[dvalley]

I don't really know pfau but have heard of him (only here on ER.org) and once tried to read one of his articles but fell asleep. However, from what you quoted the thing that jumps out at me between these quotes is that in Feb he was speaking about the 'future' not 'today' as in View#1 and #2. Again, just my guess.

 

[Finder]

I agree with your assessment.
It seems that for his February piece he was saying returns going forward will be lower, so change to 3% SWR going forward.

In his July piece he is saying that ~4% is intended to cover variations of returns, high and low over history.

On the surface it doesn't seem he is being consistent, but maybe I am missing something.

 

[Totoro]

He is not inconsistent.

The safe withdrawal rate would need to be scaled down below this, as safe withdrawal rates are more conservative than best guesses about withdrawal rates.

He is saying that the most likely scenario is 4.2%. However the "safe" scenario (whatever that means) is below that. So that's the 3% or even less.

 

[haha]

He is not inconsistent.



He is saying that the most likely scenario is 4.2%. However the "safe" scenario (whatever that means) is below that. So that's the 3% or even less.

Well, this is blindingly helpful information from him.

Likely no one would think of this on his own.
Ha

 

[Running_Man]

From what I am understanding, in this month's article he is merely looking at the historical data from 2 points of view to see what that says about withdrawl rates and reporting what their conclusions are, these are not his conclusions rather conclusions if only the data availble from the past was the data to provide future returns.

I could not access the first link but in other writings challenging the 4% rule Wade Pfau was challenging that the data set is complete enough based on the low interest rate enviroment we are in and lack of similar circumstances in the past.

 

[FIRE'd@51]

While I believe that market valuation affects the SWR, I have absolutely no faith in the 4.2% number derived from a fitted curve with an R-squared of 53%.

 

[Animorph]

While I believe that market valuation affects the SWR, I have absolutely no faith in the 4.2% number derived from a fitted curve with an R-squared of 53%.

Yep. We haven't had much luck with that sort of analysis yet. It would be interesting to see if there was anything more predictive. A simple percent down from last peak might do as well. Nonetheless, we do know we're likely to be on the wrong end of that curve, if not exactly what the SWR will be.

 

[LOL!]

3% 4% 4.2% 3.14159265%
Haven't you figured out by now that no one can predict the future with any precision?

Every time I read a sustained withdrawal rate article, I mentally put a "plus-or-minus 1.5%" after any number that the authors come up with. If you want a rock-solid number cast in stone or forged into your steel blade, then you ain't going to get it.

The answer is that you will have to adjust your spending if things go bad for you. It's that simple.

 

[haha]

3% 4% 4.2% 3.14159265%

The answer is that you will have to adjust your spending if things go bad for you. It's that simple.

The most transparent way to do this is to abjure "SWR" and use a % or previous year ending portfolio value, perhaps with some genuflection to upper and lower bounds on the withdrawal number.

Speaking of myself, I would never spend say 4% of the high valuation ending value of a boom year. Some of that would perhaps be liquidated as if I were going to spend it, but added to cash awaiting a more opportune investment climate.

The key is valuation. Everything else is derivative.

Ha

 

[LOL!]

An easy way to adjust spending it to stop buying food and eat at the free gnosh-pits in your local grocery store or Costco. Wine included.

Actually, those "best places to retire" articles should rank places based on the nearby free food sources.

 

[David1961]

3.14159265%
Haven't you figured out by now that no one can predict the future with any precision?

It's as easy as pi!

 

[Midpack]

Like a few others, Pfau has been using the A word too often for my tastes too, but he's still worth reading IMHO.

I've come to the conclusion SWR is most useful during accumulation to estimate "how much" one needs to retire, it's good for that. After spending, the user need only factor in years, risk tolerance, other income sources, etc. to decide when to (voluntarily) pull the retirement trigger.

During the retirement spending years, SWR might be a decent rough guide on what to expect, but almost all of us will have to adjust spending/withdrawals up or down (could be substantial) as the years pass, no way around it. One need only look at the basic output of FIRECALC to see that clearly - anyone think the future will be more predictable?

Articles about SWR are useful, as long as you recognize it's an academic exercise by definition. There is no "right" or wrong...

 

[W2R]

Makes me question his credibility.

And here I thought I was the only one around who was not a fan of Pfau.

 

[audreyh1]

And here I thought I was the only one around who was not a fan of Pfau.

As long as you don't live in Pharr, we're good!

 

[REWahoo]

Pharr Phau fans? Wash your mouth out at the Price Pfister, sister.

 

[aim-high]

... One need only look at the basic output of FIRECALC to see that clearly ...

Out of curiosity, what were the FIRECalc parameters you used for that graph?

 

[foxfirev5]

It's something we all need to discover for ourselves. As LBYM types I don't see much danger in this groups typical SWR.

 

[steelyman]

I'd rather discuss things with Michelle Pfeiffer.

 

[ejman]

With the extensive LBYM training most of us had to get thru to obtain the ER badge I suspect the great unwashed majority at this forum (moi included) would have absolutely no trouble adjusting the withdrawal rate way down if conditions require it. I'm probably not the only one here that basically looks at the 4% rule as THE TOP of my allowable withdrawal range.

 

[photoguy]

Maybe I'm missing something, but it appears to me Wade Pfau is putting out conflicting information about the safety of the 4% WR.

Pfau's two views aren't really conflicting but the explanation in the blog post is not very clear. In the second case, he's saying that the regression's best estimate at current PE10=26 is that the maximum sustained WR (MWR) rate is 4.2%.

The regression should be designed to predict the *average* MWR so roughly half the time the number will be higher and half the time the number will be lower.

In other words, at PE10=26, you will have a 50% failure rate if you use a WR of 4.2%.

Obviously, 50% failure rate couldn't be called safe and it's why he says "The safe withdrawal rate would need to be scaled down below this, as safe withdrawal rates are more conservative than best guesses about withdrawal rates."

I hope this clarifies things, but I'm not great at exposition either.

 

[ERD50]

....

During the retirement spending years, SWR might be a decent rough guide on what to expect, but almost all of us will have to adjust spending/withdrawals up or down (could be substantial) as the years pass, no way around it. One need only look at the basic output of FIRECALC to see that clearly - ...

I don't see how you come to this conclusion.

' almost all of us will have to adjust .... no way around it' ? With the defaults, 95% of the people historically needed no adjustments. That's not 'almost all needing to adjust', it's only a small minority who needed to adjust (historically, of course).

-ERD50

 

[Huston55]

Pfau's two views aren't really conflicting but the explanation in the blog post is not very clear. In the second case, he's saying that the regression's best estimate at current PE10=26 is that the maximum sustained WR (MWR) rate is 4.2%.

The regression should be designed to predict the *average* MWR so roughly half the time the number will be higher and half the time the number will be lower.

In other words, at PE10=26, you will have a 50% failure rate if you use a WR of 4.2%.

Obviously, 50% failure rate couldn't be called safe and it's why he says "The safe withdrawal rate would need to be scaled down below this, as safe withdrawal rates are more conservative than best guesses about withdrawal rates."

I hope this clarifies things, but I'm not great at exposition either.

Photoguy-I think this is well stated.

I am a Pfau fan. I like the data-based, analytical approach to his blog, and it contains a lot of useful links and reference info. For example, his recent posts on SWR investment approaches recommended by Kitces and Guyton, and his March 2014 post on various investment strategies were both very informative for me and great departure points for further study.

 

[clifp]

I don't see how you come to this conclusion.

' almost all of us will have to adjust .... no way around it' ? With the defaults, 95% of the people historically needed no adjustments. That's not 'almost all needing to adjust', it's only a small minority who needed to adjust (historically, of course).

-ERD50

I think will 95% will feel the need to adjust (either upwards or downwards). It is hard to imagine any Y2K retiree looking at 10+% withdrawal of their current portfolio in Jan of 2009 and thinking everything is going to be ok for the next 20 years.

It is possible that the Y2K retiree maybe still be ok in 15 years but I wouldn't bet big money it.

My starting withdrawal was under 3%, and my investments s outperformed the 75/25 benchmarks and I still cut spending in 2009.

 

[Looking4Ward]

Not sure which concerns me more - that I have a 10% chance of running out of money before I'm 90, or that I only have a 40% chance of living to 90......