Page 3 of 6

Re: PP, Retirement, and Safe Withdrawal Rates

Posted: Mon Jun 22, 2015 8:16 pm
by MachineGhost
Tyler wrote: BTW -- that's one more reason to be wary of reading too much into PP performance by substituting silver for gold.  That alone increases the PP's St. Dev to 15.98.  More than double!  That could trigger a few failures that a portfolio with gold would not.
In reality, it would be risk paritized.

Re: PP, Retirement, and Safe Withdrawal Rates

Posted: Mon Jun 22, 2015 8:31 pm
by Tyler
flagator wrote: In any event, how can the PP fulfill its role if metals are "being managed"?
If price manipulation negates a portfolio then stocks and bonds are off limits as well.  Nothing goes untouched these days.

IMHO, the PP works because it is diversified.  If people try to hold down gold to prop up the dollar, the PP holds dollars.  If the FED reduces interest rates to prop up stocks, the PP holds stocks.  What drives the change in any one asset is less important than where the money goes afterwards. 

Re: PP, Retirement, and Safe Withdrawal Rates

Posted: Wed Jun 24, 2015 2:46 am
by mathjak107
Tyler wrote:
mathjak107 wrote: the permanent portfolio can be a double edge sword as far as success rates go using the traditional  constant spending 4% rule.

the difference between bill bengans safemax results  and the trinity study is bill bengan actually had a higher success rate because he used 5 year gov bonds . it was a bit over 4.15% with no failures.

the trinity study used more volatile longer term corporate bonds and so there were failure periods.

more volatility can skew results  during unfavorable  time frames .
You're absolutely correct on the effect of volatility on portfolio failure.  That's where the PP really shines in retirement.

Since 1972:
A 50/50 Stocks/Total Bond Trinity portfolio has a 4.95% real CAGR with a 10.28 St. Dev.
A 50/50 Stocks/5-year T-bills Bengen portfolio has a 5.05% real CAGR with a 10.08 St. Dev.
The PP  has a 4.46% real CAGR with a 6.87% St. Dev.  (Use 2-year treasuries for cash and you get 4.96% CAGR, 7.09 SD)

Despite having the most volatile bonds and a big chunk of even more volatile gold, the PP has much less total volatility than the portfolios used in all of the SWR studies.  That's where Harry Browne was so insightful -- he didn't cut back on volatile assets, he loaded up on uncorrelated volatile assets that respond well in every economic condition to balance each other out.  It's the resulting low volatility for the portfolio as a whole that allows it to outperform other portfolios at the same WR regarding retirement failures even while the average return is sometimes lower.  People who only look at averages miss the measurable effect of uncertainty.

BTW -- that's one more reason to be wary of reading too much into PP performance by substituting silver for gold.  That alone increases the PP's St. Dev to 15.98.  More than double!  That could trigger a few failures that a portfolio with gold would not.
of course the problem is that while the volatility of the portfolio may be lower the wrong asset being volatile and hurting returns  can hurt the retirement .

the longer term bonds used in the trinity study were more volatile than the 5 year bonds used in bill bengan's study.  those longer term bonds made the portfolio fail in the high inflation periodss in the 1966 retiree group.

our long term treasuries are akin. they are very volatile and perhaps they too would have caused failures just like the longer term corporate bonds did.

Re: PP, Retirement, and Safe Withdrawal Rates

Posted: Wed Jun 24, 2015 10:14 am
by Kevin K.
Don't know if you've seen this recent post by Wade Pfau, but I thought it was an excellent short discussion about how much the type of bonds used in the two most influential SWR write-ups matters:

http://retirementresearcher.com/safe-wi ... i=18360981

I think Bob Clyatt's approach to a variable SWR is a good one.

Backtesting shows the PP has been a well-crafted "bunker" (Clyatt's own admiring characterization of it) in the past, but the standard selling point on it is don't look at backtesting, look at its true diversification of assets in various economic conditions. Bill Bernstein's recent "Deep Risk: How History Informs Porfolio Design" is a really interesting and provocative "riff" on the four economic conditions the PP is designed to perform well in, and among other points he shows that it makes little sense to allocate exactly equal percentages of a portfolio to risks that are anything but equally likely to occur and that have wildly different costs to insure against. The book takes maybe an hour to read (cheap Kindle download).

Re: PP, Retirement, and Safe Withdrawal Rates

Posted: Wed Jun 24, 2015 11:43 am
by Tyler
mathjak107 wrote: the longer term bonds used in the trinity study were more volatile than the 5 year bonds used in bill bengan's study.  those longer term bonds made the portfolio fail in the high inflation periodss in the 1966 retiree group.

our long term treasuries are akin. they are very volatile and perhaps they too would have caused failures just like the longer term corporate bonds did.
Counter-intuitively, the key to addressing the high volatility of the stock market is not necessarily to dial back the volatility of the other assets.  Keeping volatile stocks but muting bonds helps a bit, but only goes so far.  Another way is to balance the volatility of multiple uncorrelated assets.

But really that's just part of the equation. Stealing a great line from Craig, a usual stock/bond allocation bases diversification on correlation between asset classes. The Permanent Portfolio looks only at how the assets correlate to the economy. The problem with corporate bonds in the Trinity study may simply be that they have a habit of becoming pretty correlated with stocks in bear markets. 

You might browse some of the posts by MachineGhost. He's done some good research on risk parity in the PP.  The book covers this as well.

(Edited for clarity)

Re: PP, Retirement, and Safe Withdrawal Rates

Posted: Wed Jun 24, 2015 11:56 am
by mathjak107
Kevin K. wrote: Don't know if you've seen this recent post by Wade Pfau, but I thought it was an excellent short discussion about how much the type of bonds used in the two most influential SWR write-ups matters:

http://retirementresearcher.com/safe-wi ... i=18360981

I think Bob Clyatt's approach to a variable SWR is a good one.

Backtesting shows the PP has been a well-crafted "bunker" (Clyatt's own admiring characterization of it) in the past, but the standard selling point on it is don't look at backtesting, look at its true diversification of assets in various economic conditions. Bill Bernstein's recent "Deep Risk: How History Informs Porfolio Design" is a really interesting and provocative "riff" on the four economic conditions the PP is designed to perform well in, and among other points he shows that it makes little sense to allocate exactly equal percentages of a portfolio to risks that are anything but equally likely to occur and that have wildly different costs to insure against. The book takes maybe an hour to read (cheap Kindle download).

To me bob's variable method makes sense , we will use it set a max budget each year. it is real time and dynamic and I can see it working well for us.

Re: PP, Retirement, and Safe Withdrawal Rates

Posted: Wed Jun 24, 2015 8:10 pm
by Dmilligan
Kevin K. wrote: Backtesting shows the PP has been a well-crafted "bunker" (Clyatt's own admiring characterization of it) in the past, but the standard selling point on it is don't look at backtesting, look at its true diversification of assets in various economic conditions.
Did Clyatt ever say why he went with a slice-and-dice rather than with the PP?
Kevin K. wrote: Bill Bernstein's recent "Deep Risk: How History Informs Porfolio Design" is a really interesting and provocative "riff" on the four economic conditions the PP is designed to perform well in, and among other points he shows that it makes little sense to allocate exactly equal percentages of a portfolio to risks that are anything but equally likely to occur and that have wildly different costs to insure against. The book takes maybe an hour to read (cheap Kindle download).
Thanks for mentioning this book. I had not heard of it but grabbed it through Kindle Unlimited today and had an opportunity to read. A quick read and interesting. Very enjoyable.

Re: PP, Retirement, and Safe Withdrawal Rates

Posted: Thu Jun 25, 2015 9:47 am
by Dmilligan
Dmilligan wrote:
Kevin K. wrote: Backtesting shows the PP has been a well-crafted "bunker" (Clyatt's own admiring characterization of it) in the past, but the standard selling point on it is don't look at backtesting, look at its true diversification of assets in various economic conditions.
Did Clyatt ever say why he went with a slice-and-dice rather than with the PP?
I've answered my own question by finding this thread over on Early Retirement forums where you and Craig had an awesome conversation on the merits of slice-and-dice versus PP.

Re: PP, Retirement, and Safe Withdrawal Rates

Posted: Tue Sep 08, 2015 11:10 am
by Tyler
So this thread made me do a lot of thinking and encouraged me to fire up the spreadsheets again.  The safe withdrawal rate was a pretty simple calculation in the Trinity and Bengen studies, so what's stopping me from calculating it for the PP? 

Mission accomplished, and the results may surprise you.

First, the 60-40 portfolio to calibrate expectations. 

[img width=500]https://portfoliocharts.files.wordpress ... -rates.jpg[/img]

The light green bars are the safe withdrawal rates based on the original Bengen study methodology -- the Max WR that would have not completely run out of money during the worst rolling retirement period of a given duration.  The dark green are my own twist -- the Max WR that would have sustained the original inflation-adjusted principal during the worst rolling retirement period of a given duration.  The Safe WR is right about what you'd expect, trending towards 4%.  So the calculations are on target.

Now the PP:

[img width=500]https://portfoliocharts.files.wordpress ... -rates.jpg[/img]

Basically, the SWR of the PP is about 5%, and a 4% WR would preserve the original principal rather than draw it down at all.  Pretty surprising, huh? 

For more info on background, assumptions, and disclaimers, read this.  For an explanation of why the SWR for portfolios like the PP departs so much from the conclusions of the Trinity and Bengen studies (and the calculators built on them), read this.  And to calculate it for your own portfolio and study just how much asset allocation really does affect withdrawal rates, try out the new Withdrawal Rates calculator here

Then after all that, start brainstorming ways to ditch the day job and let your portfolio do the work for you.  ;)

Re: PP, Retirement, and Safe Withdrawal Rates

Posted: Tue Sep 08, 2015 12:39 pm
by mathjak107
only problem is  THAT IS NOT A TRUE SAFE WITHDRAWAL RATE ,  that is only a withdrawal rate that worked over a specific time frame .

to be called a safe withdrawal rate we first need to know what that term means .

the term means that the portfolio was able to stand up to not just any 30 year rolling time frames  even if not to good but specifically stood up to  the worst times frames we have had going back to 1926.

those time frames were 1929 1937 and 1965/1966.

so why those time frames and no others ?

because more modern research by michael kitces has taught us that  what made those time frames the absolute worst time frames that to date have not been duplicated  by any other date ranges is the following :


all the failed time frames had the same common denominator . they all failed to maintain at least a 2% real return average  with a 60/40 portfolio over  THE FIRST 15 YEARS OF A 30 YEAR TIME FRAME .

all the failed periods had very respectable 30 year numbers .  but their  first 15 years were terrible and no matter how good it got after the  first 15 years it was to little to late to save things .

want to know what the actual results were over the worst 30 year periods ever ?

suppose you were so unlucky to retire in one of those worst time frames ,what would your 30 year results look like :

1907 stocks returned 7.77% -- bonds 4.250-- rebalanced portfolio 7.02- - inflation 1.64--

1929 stocks 8.19% - - bonds 1.74%-- rebalanced portfolio 6.28-- inflation 1.69--

1937 stocks 10.12 - - bonds 2.13 - rebalanced portfolio -- 7.24 inflation-- 2.82

1966 stocks 10.23 - -bonds 7.85 -- rebalanced portfolio 9.56- - inflation 5.38

for comparison the 140 year average's were:

stocks 8.39--bonds 2.85%--rebalanced portfolio 6.17% inflation 2.23%

so what made those time frames the worst ? what made them the worst is the fact in every single retirement time frame the outcome of that 30 year period was determined not by what happened over the 30 years but the entire outcome was decided in the first 15 years.

so lets look at the first 15 years in those time frames determined to be the worst we ever had.

1907--- stocks minus 1.47%---- bonds minus .39%-- rebalanced minus .70% ---inflation 1.64%

1929---stocks 1.07%---bonds 1.79%---rebalanced 2.29%--inflation 1.69%

1937---stocks -- 3.45%---bonds minus 3.07%-- rebalanced 1.23%--inflation 2.82%

1966-stocks minus .13%--bonds 1.08%--rebalanced .64%-- inflation 5.38%

it is those 15 year horrible time frames that the 4% safe withdrawal rate was born out of since you had to reduce from what could have been 6.50% as a swr down to just 4% to get through those worst of times.

while 6.50% to 4% does not sound like a lot 1 million at 4% is an initial draw rate of 40k , at 6.50% you could have had 65k . that is a whopping 60% more .

so what it boils down to is any time you fall below a 2% real return average over the first 15 years you run the danger of 4% not holding. but even a 1/2% cut in spending will make you whole again over the next 15 years or longer.


starting in  the early 1970's totally misses the disasters the 4% swr was based on . if we discount them out then 6.50% would be the  draw rate for a 60/40 mix.

you cannot pull those time frames out as they are the time frames the safe withdrawal rate was created off of .  to be called a safe withdrawal rate by comparison the pp would have had to maintain at least a 2% real return over the first 15 years of specifically those dates above .

otherwise if we eliminate them the  average  withdrawal rate for a 60/40 mix is 6.50%  but we can not call it a safe withdrawal rate because like the pp it has not made it through the worst of times .

Re: PP, Retirement, and Safe Withdrawal Rates

Posted: Tue Sep 08, 2015 12:52 pm
by Pointedstick
If you define "safe withdrawal rate" to require data going back before the PP was capable of existing because the dollar was still pegged to gold, then the PP has no safe withdrawal rate and the term cannot be used. I think Tyler's data speaks for itself.

Re: PP, Retirement, and Safe Withdrawal Rates

Posted: Tue Sep 08, 2015 12:53 pm
by Tyler
Mathjak -- you need to read the links I included.  I am very open about the limitations of the research.

I used the exact same methodology of the Bengen study for the analysis, and the results for the 60/40 portfolio are basically identical.  The data is what it is.  And the reasoning behind it is very straightforward. 

Even if you do not trust the actual numbers, you must inherently realize that since the returns of your own Fidelity Insights portfolio are not the same as the results of the S&P500 index (why else would you invest differently except to get different results?) then the SWR must be different as well.  The Trinity and Bengen studies you quote did not analyze your portfolio!

I don't claim the data is a perfect representation of the SWR throughout all of history.  But it shows very clearly that asset allocation has a huge effect on SWRs.  You owe it to yourself to consider the implications for your own retirement expectations.

Re: PP, Retirement, and Safe Withdrawal Rates

Posted: Tue Sep 08, 2015 12:55 pm
by mathjak107
exactly the problem pointed stick . no one can tell what the swr on the pp would have been . it can never be tested for a true swr rate in the sense of what  the term safe withdrawal rate represents .

that has been my problem with using it for retirement day 1 . no history of what that rate actually would be in comparison .

Re: PP, Retirement, and Safe Withdrawal Rates

Posted: Tue Sep 08, 2015 12:57 pm
by mathjak107
Tyler wrote: Mathjak -- you need to read the links I included.  I am very open about the limitations of the research.

I used the exact same methodology of the Bengen study for the analysis, and the results for the 60/40 portfolio are basically identical.  The data is what it is.  And the reasoning behind it is very straightforward. 

Even if you do not trust the actual numbers, you must inherently realize that since the returns of your own Fidelity Insights portfolio are not the same as the results of the S&P500 index (why else would you invest differently except to get different results?) then the SWR must be different as well.  The Trinity and Bengen studies you quote did not analyze your portfolio!

I don't claim the data is a perfect representation of the SWR throughout all of history.  But it shows very clearly that asset allocation has a huge effect on SWRs.  You owe it to yourself to consider the implications for your own retirement expectations.

without those dates you never tested the pp for a safe withdrawal rate as compared to the work of bengan and the trinity .

you tested it for a withdrawal that met at least 4% over different time frames but not the worst case the trinity is based on .

not buying in to any thing else because the facts are the facts . if we can't test the pp over the same worst time frame all bets are off and  you are using a different set of benchmarks  which are not the worst case so of course things will look so much better .

the fact is if we can't test the pp by equal measure then we should refrain from the terminology of a safe withdrawal rate since that does not meet the definition of just what the term means .



it specifically means it weathered those dates , not your dates .



sorrry my friend you just can't measure the pp in that comparison .

Re: PP, Retirement, and Safe Withdrawal Rates

Posted: Tue Sep 08, 2015 1:07 pm
by Tyler
mathjak107 wrote: exactly the problem pointed stick . no one can twell what the swr on the pp would have been . it can never be tested for a true swr rate in the sense of what  the term safe withdrawal rate represents .

that has been my problem with using it for retirement day 1 . no history of what that rate actually would be in comparison .
Then what's the SWR for your own portfolio?  The Fidelity Insights growth fund you use started in 1987, and the growth & income fund in 1994.  Neither follows the very specific portfolios of the Trinity or Bengen studies. 

The best we can all do is to compare different portfolios over the longest timeframe possible and make a well-informed decision on how to proceed.  I'm just trying to study the problem from a new perspective, as I think many people (you included) carry a few misconceptions about SWRs and how they apply to their own portfolios. 

Re: PP, Retirement, and Safe Withdrawal Rates

Posted: Tue Sep 08, 2015 1:17 pm
by mathjak107
don't know  to be fair , so i would never call it it a safe withdrawal rate  . it should perform pretty close to as tested being 50% equity and 50% intermediate term bonds but i can't say for sure it will actually be  the as tested results . but even within 25%  of it works fine since we know the  actual 4% swr leaves more money than you started with 90% of the time over all the rolling periods . that is a pretty big awe crap fudge factor

the pp is much to radical for me to assume even being  that close . 


i have no problem with your charts showing how the pp did from the 1970's on but i do have issues with it being called a safe withdrawal rate and being compared to the actual trinity where the mixes would have much higher withdrawal rates if we eliminate those time frames .

Re: PP, Retirement, and Safe Withdrawal Rates

Posted: Tue Sep 08, 2015 1:18 pm
by Tyler
mathjak107 wrote:
the fact is if we can't test the pp by equal measure then we should refrain from the terminology of a safe withdrawal rate since that does not meet the definition of just what the term means .
Fair enough. 

The charts show the SWR since 1972.  The Trinity and Bengen studies show SWRs since 1926.  Looking at the same 60/40 portfolio over both timeframes, the SWRs are virtually identical, with the longer timeframe being a few tenths of a percent lower because of the rough historic patches you mention.  Comparing results side by side over the same timeframe, you see that results can vary widely based on asset allocation.  This factually must be the same for the longer timeframe as well, but we don't (yet) have the data to compare numbers since 1926 directly. 

Those are the facts.  You are free to draw your own conclusions.

Re: PP, Retirement, and Safe Withdrawal Rates

Posted: Tue Sep 08, 2015 1:19 pm
by sophie
One thing that always bothered me about a safe withdrawal rate expressed as a fixed percentage, is that it assumes your spending is going to increase or decrease along with portfolio value.  In practice, there is a "floor" of spending that won't decrease (property taxes, rent, mortgage, home repairs, food, utilities, medications etc).  Conversely, spending doesn't necessarily increase when the portfolio value goes up.

We had a thread a while back where portfolio sustainability was calculated by determining nominal portfolio value after an annual cash withdrawal that started at a fixed amount (3%, 4%, or 5% of the portfolio) and then increased according to inflation.  While that's still somewhat idealized, it's far more useful a representation I think.  Perhaps injecting random large expenses on top of a basic living "stipend" would be yet more realistic.  Like, the house needs a new roof or a new heating system, you decide to take an expensive vacation for your 50th wedding anniversary, a major medical expense comes up, etc.

Of course I'm suggesting this for the Tylers and Pet Hogs of the world, whose contributions have been amazing and incredibly appreciated.  I'm afraid I don't have time to monkey with spreadsheets for fun anymore.

Re: PP, Retirement, and Safe Withdrawal Rates

Posted: Tue Sep 08, 2015 1:24 pm
by Tyler
Good point, Sophie.

To be clear, my calculations assume that the withdrawal rate quoted is the percentage of the portfolio in year one.  After that, it is adjusted by inflation and not by portfolio value.  I agree that this is the better method.  And you're also correct that it still doesn't tell the whole story, and that simply being smart about expenses rather than mechanically spending the amount of money the spreadsheet tells you to can massively improve your retirement performance. 

Re: PP, Retirement, and Safe Withdrawal Rates

Posted: Tue Sep 08, 2015 1:30 pm
by Pointedstick
Awesome tool, Tyler. Your site is getting better and better.

It doesn't surprise me in the least that the PP has historically been able to sustain 5% withdrawals without depleting the principal; my own calculations a few years ago were in the same ballpark.

My current favorite PP alternative (20% midcaps 20% SCV 10% EM 10% gold 20% LTGB 20% short-term treasuries) looks like it would sustain almost 6% withdrawals without depleting the principal. More gold would push it over 6%.

I also think Sophie's point about expenses being unpredictable is a very salient one. My own expenses have been either flat or declining over the past couple of years as I have learned to do more and more myself rather than paying other people to do them for me at the wildly inflated prices that are necessary to sustain oneself in many fields. Plumbers here charge north of $100 an hour for labor alone!

It might be useful to have options for determining spending, or even adding of subtracting from inflation to account for people who don't trust CPI. If it could be shown that a given portfolio would survive even if CPI were only half of the supposed real value, for example, that would be a very powerful piece of information to have.

Re: PP, Retirement, and Safe Withdrawal Rates

Posted: Tue Sep 08, 2015 1:34 pm
by Pointedstick
Also FYI Tyler, the new calculator appears to occasionally download empty files named "chartiframe" (zero KB, no content) to the user's specified download location. It's happened twice so far.

Re: PP, Retirement, and Safe Withdrawal Rates

Posted: Tue Sep 08, 2015 2:25 pm
by mathjak107
don't forget though that the 4% swr rate is only the income side of things . what varies is what is left in the bucket at the end .

it can range from very little as in the worst of times to  3x what you started with in the best of times .

so you can't just compare draw rates but what is actually left at the end  too .

historically using a 60/40 mix there has been more than you started with 90% of the time and more than 2x what you started with 67% of the time . median wealth has been 2.8x what you started with .

was that reflected in your calculations ?


remember the whole idea of a safe withdrawal rate was the income held constant through thick and thin  and the markets effected the balance at the end without running dry . but since the 4% swr is based on the worst of times and we have never duplicated the worst of times since 1966 the 4% swr as is , has left way to much money left over .

Re: PP, Retirement, and Safe Withdrawal Rates

Posted: Tue Sep 08, 2015 2:54 pm
by Tyler
@Mathjak -- That's true.

I used the Bengen SAFEMAX methodology that mapped the maximum WR for the single worst retirement period with no regard for percentages. The SWR shown is the worst case, and all other retirement periods will be better.  I accounted for end portfolio values using the Sustainable WR.  I personally prefer that to getting bogged down in arguments like "is a 50% chance acceptable, or is 90% better?" but I know where you're coming from.  It's true that a single SWR alone does not tell the whole story when it comes to end portfolio values.

Re: PP, Retirement, and Safe Withdrawal Rates

Posted: Tue Sep 08, 2015 2:59 pm
by mathjak107
we both agree there .

i think trying to compare  the pp  when it comes to swr is very difficult on so many levels .  trying to compare the pp's 5% with a 4% true swr based on far worse time frames and no regard for what is left as a balance really does not compare much at all .

in fact remember that bill  in safemax had better results than the trinity did because  safemax used 5 year gov't  bonds while the trinity used intermediate corporate bonds .  the GREATER VOLATILITY  in the portfolio of the corporate longer term bonds actually hurt things  and the trinity had more failure periods . 

so we really don't know what the pp will do vs conventional mixes , in fact in the new normal neither may conform to what was so basing results on conditions that no longer exist in both cases may be the wrong thing to do .

Re: PP, Retirement, and Safe Withdrawal Rates

Posted: Tue Sep 08, 2015 3:01 pm
by sophie
Pointedstick wrote: It might be useful to have options for determining spending, or even adding of subtracting from inflation to account for people who don't trust CPI. If it could be shown that a given portfolio would survive even if CPI were only half of the supposed real value, for example, that would be a very powerful piece of information to have.
+1.

The more I think about it, the more I think CPI is inadequate for capturing year to year increases in retirement spending.  It doesn't effectively capture two of the largest expenditures:  property taxes and medical costs - the former is not considered at all, and the latter is considered separately (as far as I can tell from the BLS website).  Also, as you get older you may need more and more services for things you used to do yourself.