Dr. Pfau looks at the 4% rule in an interview

[kevink]

Thanks all for the thoughtful comments. It's interesting that both FIRECalc and Dr. Pfau use past performance of a couple of key asset classes to base their predictions on, but Pfau's arguments are based on a prediction that future stock and bond market returns will depart greatly from the past. Essentially his advice is "all in" on equities, with an annuity to make it possible to sleep at night. Given what happened in 2008 I personally wouldn't feel comfortable with that kind of exposure if I were in my late 60's or 70's.

We were invested in Clyatt's slice-and-dice RIP Portfolio during the market crash and had paper losses of around 21%. That prompted a search for a more robust defensive approach, and I've been invested in Harry Browne's Permanent Portfolio ever since. I mention it not to "sell" it - knowing it's far too out there for most here - but because it does seem to provide another interesting way to effectively build your own annuity, if you will, providing quite consistently 4% real returns without the counterparty risks of annuitiies. There are some thoughts on this on one of the specialized PP boards:

Viewing the PP as an Annuity - Page 1

Midpack I am in awe of your FIRECalc savvy but would only say that given my risk-averse nature I'd go with the 95% rule which by definition guarentees never running out of money vs. relying on FIRECalc projections based on past performance of two asset classes under circumstances that Pfau and others are saying are gone for good. For all of our sakes I hope we are on the verge of another bull run in the markets, but the older I get the more I tend to live by the old Will Rogers saw: "I'm more concerned about the return of my capital than the return on it."

 

[samclem]

• On the Investigate tab, I chose the Given a success rate, determine spending level for a set portfolio, or portfolio for a set spending level, left success rate at the default 95%, and chose Spending Level. I am guessing this is where our inputs differed.

If that misrepresents the base case, please let me know so I can correct it.

Here's what I did:
START HERE:
Spending: $34,300 (reversed-in from your numbers in Case 1. This, I think, sets 3.43% as the amount to be withdrawn each year once we choose the "Percent of Remaining Portfolio" spending model.)
Portfolio: $1,000,000
Years: 35
Submit
OTHER INCOME/SPENDING: Use defaults
SPENDING MODELS: Select Percent of Remaining Portfolio. Leave the "Clyatt box" at 0%. Submit.
YOUR PORTFOLIO: Use "Total Market (FIRECalc default). ER is kept at .18% (BTW, changing this to 2-3% is probably the easiest way to "fool" FIRECALC into using lower annual returns to simulate the scenario Dr Pfau is suggesting) Submit.
Don't go the "investigate" tab, just look at the results screen. Along with the depictions is this text:

In other models in FIRECalc, "failure" means the portfolio drops to zero. Since you are limiting spending to a percentage of your remaining portfolio, the total balance should never reach zero — but it could become pretty small in some situations. Pay attention to the spending graph, below. Since we can't use portfolio failure as a metric, FIRECalc is following the lead of the 95% Rule from Work Less, Live More, in which one of the goals is for the portfolio to be as big (after adjustment for inflation) at the end of the 35 years as it was when you started. FIRECalc found that 98.1% of the time, the portfolio you would have left behind exceeded the portfolio you started with.

So, now I see that is the definition of "failure": an inflation-adjusted portfolio value at the end of the time window that is less than the starting amount. That, to me, is a lot different than the "you are totally broke" failure criteria used for the "fixed spending" model. If I were 87 years old today (35 years from now) and had $10 less than the portfolio size I have right now (and the same inflation-adjusted amount I started with), it would be a "failure" under these criteria. Huh? I'd feel like I was in darn good shape--it's only gotta last a few more years and I have 25x my historic "regular" WR!

For whatever reason, after I go to the "investigate" tab I can't get this same "results screen" to show again. I have to totally exit FIRECalc and re-enter my data if I want to see this screen again.

kevink did say he wanted to use Bob Clyatt's 95% rule, so your method is probably more appropriate than what I did. It looks like the "success rate" goes from about 95% to 98% if the "95% of last year's withdrawal" rule isn't used.

 

[samclem]

DANG! Mistake in previous posting: I'd left the "inflation button" clicked on "CPI" instead of a straight 3%. Using 3% inflation the "success rate" was 69.8%, not 98% as previously stated. Anyway, the observations concerning "failure" for this situation remain the same: Is it really a "failure" if someone is 87 YO and his portfolio has the same inflation-adjusted value as when he was 52? I don't think so. I'd say a "failure" when using the "X% of annual balance" withdrawal method should be when the portfolio can't support some user-defined percentage of the initial withdrawal value. For our family, I'd put failure at about 50% of our initial withdrawal amount (adjusted for inflation). That's "move to the trailer at the fish camp" time (not that there's ANYTHING wrong with that.).

 

[audreyh1]

Wow, SamClem, thanks for the digging! Things were not making sense to me either.

I agree that 50% of the initial value, inflation adjusted, sounds like a more reasonable "failure" criteria - yet still way better than going to $0!

So IMO these two methods (% of remaining portfolio and inflation adjusted withdrawals) aren't comparable the way FIRECALC is set up. You can't say that one gives you a higher initial spending amount in comparison when it has a terminal value of 0, and the other has a terminal value of just under the starting portfolio inflation adjusted.

Personally, I would be delighted if my terminal portfolio were the same as I started with, adjusted for inflation. If it's much higher than that - I wasn't spending enough!

 

[audreyh1]

DANG! Mistake in previous posting: I'd left the "inflation button" clicked on "CPI" instead of a straight 3%. Using 3% inflation the "success rate" was 69.8%, not 98% as previously stated. Anyway, the observations concerning "failure" for this situation remain the same: Is it really a "failure" if someone is 87 YO and his portfolio has the same inflation-adjusted value as when he was 52? I don't think so. I'd say a "failure" when using the "X% of annual balance" withdrawal method should be when the portfolio can't support some user-defined percentage of the initial withdrawal value. For our family, I'd put failure at about 50% of our initial withdrawal amount (adjusted for inflation). That's "move to the trailer at the fish camp" time (not that there's ANYTHING wrong with that.).

Since CPI goes along with the historical data, what's wrong with using the historical CPI?

 

[samclem]

Since CPI goes along with the historical data, what's wrong with using the historical CPI?

Historical CPI is normally what I use, but (for whatever reason) I thought Midpack had used a straight 3% and I wanted an apples-apples comparison.

 

[samclem]

We were invested in Clyatt's slice-and-dice RIP Portfolio during the market crash and had paper losses of around 21%. That prompted a search for a more robust defensive approach, and I've been invested in Harry Browne's Permanent Portfolio ever since.

Well, everyone should do what they are comfortable with. The Permanent Portfolio has been through some changes--for a "permanent" portfolio. Harry suggested 25% physical gold, 25% equities, 25% LT Treasuries, and 25% cash. For someone looking at a 30 year horizon, I think this thread and other contemporaneous ones point out the historical advantages of having more than 25% in equities. Having an ownership stake in entities that are productive is, over time, a winner. As far as LT Treasuries--I would have a hard time making a case for them today.
But, the PP might do very well in the future. We'l know in a few decades which approach was best!

 

[mathjak107]

you can still do okay long term even with the long term treasuries in th PP as it is a given they will fall but the rebalancing over time though will eventually have them back on the job.

imagine you bought the gold portion of the permanent portfolio at the worst point in time back in the 1980's at the peak price of 800 bucks.

did you know that just rebalncing the gold with the rest of the permanent portfolio all those years had gold out performing the s&p by a fraction of a percent today . i think it was 9.2 cagr for the s&p vs 9.8 for gold..

 

[urn2bfree]

you can still do okay long term even with the long term treasuries in th PP as it is a given they will fall but the rebalancing over time though will eventually have them back on the job.

imagine you bought the gold portion of the permanent portfolio at the worst point in time back in the 1980's at the peak price of 800 bucks.

did you know that just rebalncing the gold with the rest of the permanent portfolio all those years had gold out performing the s&p by a fraction of a percent today . i think it was 9.2 cagr for the s&p vs 9.8 for gold..

But Gold gains are taxed at 28% by law (federal) ..each state may or may not add to that. Cap gains could be taxed from 0 to 15% max.

 

[mathjak107]

That is true but so is bond interest and reit dividends.

 

[Midpack]

So IMO these two methods (% of remaining portfolio and inflation adjusted withdrawals) aren't comparable the way FIRECALC is set up. You can't say that one gives you a higher initial spending amount in comparison when it has a terminal value of 0, and the other has a terminal value of just under the starting portfolio inflation adjusted.

FIRECALC reasonably represents the two approaches in terms of what income the retiree is signing up for, but you're right there's a big difference in portfolio residual under worst case, so thanks for pointing that out!

OTOH, for best case the residual for inflation adjusted withdrawals kills % or remaining portfolio and annual incomes change drastically in favor of the latter. Never easy...

 

[audreyh1]

OTOH, for best case the residual for inflation adjusted withdrawals kills % of remaining portfolio and annual incomes change drastically in favor of the latter. Never easy...

Definitely - in the best case of inflation adjusted withdrawals, you are leaving a huge amount behind. Not an optimal outcome, IMO.

That's one reason (actually, the primary reason) I chose % remaining portfolio over inflation adjusted from initial portfolio value. It's a much better way to manage leaving excess behind, and well worth the variability in annual income IMO.

I just don't understand the focus on avoiding variability in annual income, when the easiest thing in the world is to "income smooth" over several years by simply not spending all the increase after a good market year and setting some of it aside to help meet the budget after a bad market year.

 

[audreyh1]

FIRECALC reasonably represents the two approaches in terms of what income the retiree is signing up for, but you're right there's a big difference in portfolio residual under worst case, so thanks for pointing that out!

I actually don't think it's a reasonable comparison.

I didn't "sign up" for

Since we can't use portfolio failure as a metric, FIRECalc is following the lead of the 95% Rule from Work Less, Live More, in which one of the goals is for the portfolio to be as big (after adjustment for inflation) at the end of the 35 years as it was when you started. FIRECalc found that 98.1% of the time, the portfolio you would have left behind exceeded the portfolio you started with.

when I chose my method. I also don't use Bob Clyatt's 95% rule - that is a different scenario than the pure fixed % of remaining portfolio method.

The only reasonable comparison between the two is to also apply the "portfolio to be as big (after adjustment for inflation) at the end of the 35 years as it was when you started" criteria to the constant inflation-adjusted withdrawals model. Then you will have a true comparison.

 

[kevink]

Well, everyone should do what they are comfortable with. The Permanent Portfolio has been through some changes--for a "permanent" portfolio. Harry suggested 25% physical gold, 25% equities, 25% LT Treasuries, and 25% cash. For someone looking at a 30 year horizon, I think this thread and other contemporaneous ones point out the historical advantages of having more than 25% in equities. Having an ownership stake in entities that are productive is, over time, a winner. As far as LT Treasuries--I would have a hard time making a case for them today.
But, the PP might do very well in the future. We'l know in a few decades which approach was best!

IMHO basing your expectations on historical returns rather than the correlation of the asset classes you hold with economics conditions is the danger. A 60:40 stocks:bonds allocation is basically no safer than 100% equities, as this paper shows:

http://www.advisorperspectives.com/newsletters12/pdfs/Why_a_60-40_Portfolio_isnt_Diversified.pdf

This is exactly what happened in the '08 crash, when "non-correlated" assets (not just stocks and bonds, but foreign and EM stocks, U.S. bonds and emerging market, etc.) tanked in unison (while the PP in breaking even outfperformed every portfolio I know of). I don't think there's any "holy grail" allocation, but a five decade return that's on par with straight S & P 500 with a tiny fraction of the volatility ain't bad.

 

[samclem]

you can still do okay long term even with the long term treasuries in th PP as it is a given they will fall but the rebalancing over time though will eventually have them back on the job.

For many PP fans, it's a bit of a religion. Where did the 25% allocation to each come from.. why not 18%, 21%, 38%, 23%? Id the "strength" of the "savior asset" in its favored economic phase really identical to the others? And does the US economy spend the same amount of time in each one of these phases (or alternatively, is the magnitude of the activity in one type of cycle so strong as to make up for a relatively short duration)? Anyway, like all investors who are just trying to make progress in the face of an uncertain future, I wish them luck. And, I do share many of their concerns.

 

[mathjak107]

The 25% came from lots and lots of back testing by terry coxen and harry brown.

They played and played until they could find an allocation that worked well and was simple with simple to implement being a big factor.

 

[rayvt]

And by sheer coincidence, 25% happens to be one-forth of 100%.

A naive person might have just arbitrarily said, "4 holdings, so divide the total portfolio by 4 to get 25% for each holding." Stupid naif! It really takes lots and lots of high-powered backtesting to decide that the optimal weighting is 25%.

I bet that it's most likely that Coxen & Brown did it like a Chicago street repair crew does a task ---- took 2 seconds to write it down, and then nipped off behind a tree for a 6 hour nap.

 

[photoguy]

A 60:40 stocks:bonds allocation is basically no safer than 100% equities, as this paper shows:

http://www.advisorperspectives.com/newsletters12/pdfs/Why_a_60-40_Portfolio_isnt_Diversified.pdf

I just read through the first bit of the paper and his argument doesn't make any sense. In fact, I think the author is being intentionally misleading with statistics.

He claims that 60-40 is no better than a 100% equity from a risk perspective because "the correlation between 60/40 and 100% in equities is 98%". The author conveniently forgets that correlation only measures whether two variables move up or down together but it does not measure the magnitude.

So in a 50% bear market a 60-40 still might leave you with 70-80% of your initial stakes, whereas with a 100% equities you'd be down to 50%. I'd say that's a huge difference in risk.

Edit: the paper appears to have other problems like estimating tips returns when it wasn't available using a "proprietary methodology". Proprietary = Non reproducible = junk/pseudo science

 

[Midpack]

Definitely - in the best case of inflation adjusted withdrawals, you are leaving a huge amount behind. Not an optimal outcome, IMO.

That's one reason (actually, the primary reason) I chose % remaining portfolio over inflation adjusted from initial portfolio value. It's a much better way to manage leaving excess behind, and well worth the variability in annual income IMO.

I just don't understand the focus on avoiding variability in annual income, when the easiest thing in the world is to "income smooth" over several years by simply not spending all the increase after a good market year and setting some of it aside to help meet the budget after a bad market year.

I've only recently retired and DW is still working, so we haven't really committed to anything re: income/withdrawals. While I've used SWR to determine FI plus a safety factor we're comfy with, even if we use inflation adjusted withdrawals at the outset, dividends will probably be more than enough to provide income for the first years/decade, so it should be a moot point for a while. At any point along the way we could go to % of remaining portfolio, and suspect we will midway through retirement. And if all goes well somewhere out at 75-80 yo, I suspect we'll annuitize (SPIA) some income to further reduce risk and "manage leaving excess" $ behind to a minimum.

And you make a good point regarding income smoothing despite the variability of income using % remaining withdrawals. Our natural instinct will probably be to underspend no matter what our withdrawal scheme is, as it has been all our lives.

Fortunately (and unfortunately) we have lots of time to consider our options...

You're a great contributor Audrey, along with many others here. Thanks...

 

[mathjak107]

And by sheer coincidence, 25% happens to be one-forth of 100%.

A naive person might have just arbitrarily said, "4 holdings, so divide the total portfolio by 4 to get 25% for each holding." Stupid naif! It really takes lots and lots of high-powered backtesting to decide that the optimal weighting is 25%.

I bet that it's most likely that Coxen & Brown did it like a Chicago street repair crew does a task ---- took 2 seconds to write it down, and then nipped off behind a tree for a 6 hour nap.

actually there was quite a bit of testing as the thought at that time was the gold allocation was to high.

it was tested with commodities instead of gold, short term bonds instead of cash and as the fund itself has done gone off into other asset classes like swiss franks , real estate and silver.

all were dropped from the origonal concept and the 4 remained in equal amounts as the best mix as well as simple..

but the basic 4 still seem to be the best and the simplest way to go. the other asset classes only cloud things.

sometimes they will act in a scenerio precictably and other times not.

look at 2008-2009 gold broke new highs while silver plunged along with most commodities.

it was never designed for growth at all. it was only for protection against devastation. the fact they hit the lottary and gold and long term treasuries did so well was a bonus but that kind of return was never planned on.

 

[kevink]

Thanks everyone for your comments (Audrey and Midpack especially!). I did not and do not intend to "hijack" this thread into a discussion of the Permanent Portfolio, but since I perhaps foolishly brought it up I'll provide links to the two best places to learn more in case of interest. The approach has been more thoroughly scrutinized than any other portfolio I know of - understandably so given its weird asset holdings. The first link is to the current web board devoted to the PP, the other to what I believe to be the longest thread in the history of Vanguard Diehards that gave birth to not only the other forum but also the excellent recent book on the PP (link on first forum).

Clearly there are any number of portfolios that will typically yield similar or better returns, from 100% Wellesley to sophisticated slice-and-dice allocations and many others. As Bob Clyatt once said the PP is a "bunker" and as such will always be the choice of a small minority of investors.

I really appreciate Dr. Pfau's writing and research and the thoughtful discussion of it here even more. I definitely see a place for an annuity in our future and that's something that wouldn't even have crossed my mind before. Thanks all and again apologies for the PP digression.

Permanent Portfolio Discussion Forum - Index

Bogleheads • View topic - Updated Modification of Harry Browne Permanent Portfolio

 

[WestLake]

Yes, we are both 56 now, and our youngest child finished college last year. With the college tuition gone, our expenses just got a big reduction. Right on schedule per Bernicke's model!

Both of us stopped craving for "stuff" long ago. Things like new cars, clothing, fancy furniture, a bigger home, we care for less and less. And it's not even because we have "done that". Even things we have not owned or done, we simply do not care for anymore.

The only thing left that I still like to do is to travel, and I have been thinking I'd better do more of that soon, because in 10 or even 5 years, I may not even care for that.

Scary for a 56-yr old to say that, but then I am a geezer before my time.

Sure! That's why I have run Bernicke's model in FIRECalc to see how high I could go, then chuckled and went back to 3.5%WR. And that 3.5%WR may be going down with time too, because I may not care to spend. Isn't that sad?

I actually do not spend too much time thinking about LTC. I do not know why, but have a feeling that when my time comes, I will not linger long.

Thanks for your well wish. I do not know about traveling until the age of 110, but I can squeeze in a bit in the next 10 years, I will be happy and call my life a completely satisfactory one.

Mirrors my thoughts. I'm turning 55 in March and did a lot of travel in my late 30s and 40s. Scares me that my travel desires are diminishing, as that was the thing I was looking forward to in retirement. Still have the goal of buying a Porsche, but I think I am losing my taste for that also.

 

[rayvt]

buying a Porsche, but I think I am losing my taste for that also.

For me it was a BMW Z4. Then we realized that someone was going to be buying a convertible sports car with my money. The only question was if it would be me or my kids.

 

[Snidely Whiplash]

... someone was going to be buying a convertible sports car with my money. The only question was if it would be me or my kids.

Just wanted to say Thanks for your having posted this nugget of wisdom. I read this post several days ago and it has really hit home with me; This comment has been stuck in my head. I was struggling with some guilt over buying a new vehicle vs. getting something used (which has never worked out for me) and this struck a chord. I might have been confusing frugality with over-depriving myself. Your comment has put things in a bit better perspective for me.

Seriously, Thank You!

 

[Koolau]

Just wanted to say Thanks for your having posted this nugget of wisdom. I read this post several days ago and it has really hit home with me; This comment has been stuck in my head. I was struggling with some guilt over buying a new vehicle vs. getting something used (which has never worked out for me) and this struck a chord. I might have been confusing frugality with over-depriving myself. Your comment has put things in a bit better perspective for me.

Seriously, Thank You!

I appreciate your sentiment, too, Snidely Whiplash. We raised our kids to be independent. We made sure they had college educations available to them (one chose not to take full advantage, but is very independent - having started several businesses already.) My point is that we don't plan to leave a whole lot to our kids (we do hope to help them along from time to time while we are still living). So we don't plan to leave this world with any guilt feelings - at least not financial guilt feelings - toward the kids (or anyone else.)

By the way, I love your name! I always loved the "villain" names in Rocky and Bullwinkle.