Is the successful salaried retail investor a myth?

[Libertarian666]

NS, I was forced to retire because of my age (56) and hate it. Don't rush into it until you're ready.

Damn, that's young to be forced to retire. What field were you in? I'm 64 and work in software development. I can't do the 16-hour days cranking out code like I used to but I can still accomplish more in 8 hours than any young whippersnapper I've yet seen so I'm not really worried about that happening to me. Not yet, any way.

Same age and field here. However, I'm not too worried about finding things to do after retirement (whenever that is). To begin with, I have to update my C++ tutorial book and propose one of my libraries as an addition to the standard library. And then there are the Rosetta Stone Spanish levels 3-5 to tackle.

[Tyler]

Alright.  Now this is getting interesting.  I managed to automate the PP rebalancing in Excel, which allowed me to start doing some more significant number crunching. 

Since retirement portfolio success is highly dependent on the year you retire (if the market crashes 50% in year 1, your WR just doubled), I wanted to determine the maximum withdrawal rate that the Permanent Portfolio can support in retirement regardless of what year you retire.  My definition of success is different than retirement calculators like Firecalc -- I want the value of the portfolio not simply to remain positive, but to also remain constant in inflation-adjusted dollars to your initial investment.  I also used the actual PP rebalancing methodology, with 15/35 bands and taking expenses every year out of the cash component.  Finally, the withdrawal rates shown are adjusted to the CPI every year so your spending power is constant. 

To find the optimal WR, I ran theoretical 10-year retirement scenarios in every rolling decade from 1972-2012 using Craig's data pulled from WiseOne's spreadsheet.  I used 10-year intervals rather than typical 30-year retirement intervals not only to get more data points, but also because my goal was to preserve my initial investment.  I figured that if I could identify the max WR that could survive the worst decade 100% intact, it would most likely last forever.

I assumed in year zero for each scenario the PP was perfectly balanced at 25%ea.  Rebalancing events were calculated automatically, and I then adjusted the initial withdrawal rate until the final PP value after ten years (including drawdowns for expenses) matched the inflation adjusted value of the initial portfolio.  The results are as follows:



The average "Tyler" SWR for the PP historically is a pretty amazing 5.3%.  The best year to retire was 1978, where one could have supported a ridiculous 7.7%WR for ten years without depleting capital at all.  The worst year to retire was 1987 where the max WR was still a healthy 3.8%.  I haven't run a full analysis of a 50/50 BH portfolio yet, but I can tell you that using the same methodology the max SWR is -2.5% if you retired in 1972.  That's not a typo -- if you didn't add at least 2.5% to your portfolio every year, you would have lost value ten years later.  That's why long-term SWR discussions for stock/bond mixes usually center around very conservative numbers.

Frankly (assuming I haven't screwed up somewhere, of course) I find this astonishing and very reassuring.  Basically, from what I can tell the conservative SWR for the Permanent Portfolio to last indefinitely is right around 4%.  And an older retiree with less worry about maintaining principal could easily push to 5% without much argument from me.  Compare this to most SWR calculations for "traditional" portfolios (where 4% may only sustain you for 30 years before you're broke), and the PP is one great retirement portfolio.

[MediumTex]

Nice work Tyler.  That's good stuff.

The PP is like a brick house.

[dragoncar]

Nice work Tyler.  That's good stuff.

The PP is like a brick house.

Mighty mighty, just lettin' it all hang out

[D1984]

The PP is like a brick house.

Unless one happens to have the misfortune to be invested in it when two volatile assets are declining at once over a fairly long period and the other asset plus the cash isn't enough to offset them. Ask any Japanese PP investor from 1990-2002 how "safe" his/her portfolio was and what the sustainable withdrawal rate would have been to keep from depleting any capital value after inflation (hint...it would be a NEGATIVE SWR as far as I can tell).

[Tortoise]

So does the PP allow a surprisingly high SWR primarily because of (a) its 25% cash allocation or (b) the low volatility of the overall portfolio? It sounds like it's (a), but I'm just curious as to whether anyone here disagrees with that assessment.

An SWR comparison analysis between a 3x33% PP (no cash) and a traditional 4x25% PP would be fascinating to see.

[koekebakker]

The PP is like a brick house.

Unless one happens to have the misfortune to be invested in it when two volatile assets are declining at once over a fairly long period and the other asset plus the cash isn't enough to offset them. Ask any Japanese PP investor from 1990-2002 how "safe" his/her portfolio was and what the sustainable withdrawal rate would have been to keep from depleting any capital value after inflation (hint...it would be a NEGATIVE SWR as far as I can tell).

That's a long period of negative returns. Problem is: there's no strategy that would have done much better.

The PP will  always be struggling when interest rates are really low. If gold and stocks drop at the same time in a low interest rate environment things could get really, really ugly.

[WiseOne]

So does the PP allow a surprisingly high SWR primarily because of (a) its 25% cash allocation or (b) the low volatility of the overall portfolio? It sounds like it's (a), but I'm just curious as to whether anyone here disagrees with that assessment.

An SWR comparison analysis between a 3x33% PP (no cash) and a traditional 4x25% PP would be fascinating to see.

Probably both.  The cash is the reason why the portfolio was able to ride out the losing years - there were no rebalances in years where real return was worse than -0.3%.  And the duration of losing years is limited by the portfolio's stability.

Tyler, how did you automate rebalancing??  Would love to learn a new Excel trick.

[stone]

The PP is like a brick house.

Unless one happens to have the misfortune to be invested in it when two volatile assets are declining at once over a fairly long period and the other asset plus the cash isn't enough to offset them. Ask any Japanese PP investor from 1990-2002 how "safe" his/her portfolio was and what the sustainable withdrawal rate would have been to keep from depleting any capital value after inflation (hint...it would be a NEGATIVE SWR as far as I can tell).

I hoped that the Japanese long term treasuries would have saved the day? The yield fell from 7% to 0.7% didn't it? Also deflation would have to be added. So each deflationary year the draw down would have amounted to a smaller nominal amount.

[CA PP]

should we use a data set of real returns instead of nominal returns?

i did an analysis a while back with r-returns for a CA PP and got about 3.5% r-swr the portfolio real value to to decrease.

i guess you re at about the same with 5.5% nominal swr.  it did however made me realise that i would get less real dollars than one woujld expect looking at nominal figures.

[Pointedstick]

Believe it or not, CA PP, those are real returns.

[Tyler]

So does the PP allow a surprisingly high SWR primarily because of (a) its 25% cash allocation or (b) the low volatility of the overall portfolio? It sounds like it's (a), but I'm just curious as to whether anyone here disagrees with that assessment.

An SWR comparison analysis between a 3x33% PP (no cash) and a traditional 4x25% PP would be fascinating to see.

It's hard to say, but I suspect it's a combination of a, b, and c) the rebalancing bands that allow for multi-year momentum runs to pad the winning asset.  I'm still looking at this, but I have noticed that over long periods of time the bands/cashDD method does diverge from the annual rebalance/(sell 4% of everything) method by quite a bit, especially when you creep towards very high withdrawal rates that stress the system. 

[Tyler]

Tyler, how did you automate rebalancing??  Would love to learn a new Excel trick.

1) I created a new column that looked for portfolio percentages outside the bands:

=IF(OR(R12>0.35, R12<0.15, S12>0.35, S12<0.15,T12>0.35,T12<0.15,U12>0.35,U12<0.15),1,0)

This returns a 1 if true and a 0 if false

2) I used an IF statement in the interest calculation field that looked at the number in #1.  If it's a 1, it does the calculation using the previous total portfolio divided by 4.  If it's a 0, it calculates normally.

=IF($V17=1,$P17/4,M17)*(1+C18/100)

[Tyler]

should we use a data set of real returns instead of nominal returns?

i did an analysis a while back with r-returns for a CA PP and got about 3.5% r-swr the portfolio real value to to decrease.

i guess you re at about the same with 5.5% nominal swr.  it did however made me realise that i would get less real dollars than one woujld expect looking at nominal figures.

The data I used had the nominal returns for the PP results every year, but I also adjusted overall calculations for the real-life CPI-U data in the same year.  So the results do show real returns. 

Any time you do calculations like this there are always sources of error between different people.  I think one of the biggest (and most common) is to assume a constant inflation rate or a smooth CAGR every year when the real world doesn't act that way.But even simple things like using a different inflation metric may change the results.

Despite all the numbers, never confuse precision with accuracy.  I still believe the best retirement strategy requires personal flexibility rather than faith in spreadsheets.

[stone]

So does the PP allow a surprisingly high SWR primarily because of (a) its 25% cash allocation or (b) the low volatility of the overall portfolio? It sounds like it's (a), but I'm just curious as to whether anyone here disagrees with that assessment.

An SWR comparison analysis between a 3x33% PP (no cash) and a traditional 4x25% PP would be fascinating to see.

It's hard to say, but I suspect it's a combination of a, b, and c) the rebalancing bands that allow for multi-year momentum runs to pad the winning asset.  I'm still looking at this, but I have noticed that over long periods of time the bands/cashDD method does diverge from the annual rebalance/(sell 4% of everything) method by quite a bit, especially when you creep towards very high withdrawal rates that stress the system. 

Is some of the effect actually due to the even draw down approach in general having more cash (back to 25% each year) whilst the kosher method hovers around 15% cash as the drawdowns from cash trigger rebalances? If so, it might actually bode well for a 3x33%PP (no cash) where draw downs came from whichever asset was leading???

[Tyler]

Is some of the effect actually due to the even draw down approach in general having more cash (back to 25% each year) whilst the kosher method hovers around 15% cash as the drawdowns from cash trigger rebalances? If so, it might actually bode well for a 3x33%PP (no cash) where draw downs came from whichever asset was leading???

I dunno - Could be.  Deserves some study. 

I haven't looked at that extensively because generally whenever I've tried to tweak the PP is doesn't work. 

[dragoncar]

So does the PP allow a surprisingly high SWR primarily because of (a) its 25% cash allocation or (b) the low volatility of the overall portfolio? It sounds like it's (a), but I'm just curious as to whether anyone here disagrees with that assessment.

An SWR comparison analysis between a 3x33% PP (no cash) and a traditional 4x25% PP would be fascinating to see.

I modified the spreadsheet for 30/30/30/10 with 42/18 bands except for cash, which had a 0/15 band (but the upper band was irrelevant).  The end value was higher than the 25x4.

[D1984]

The PP is like a brick house.

Unless one happens to have the misfortune to be invested in it when two volatile assets are declining at once over a fairly long period and the other asset plus the cash isn't enough to offset them. Ask any Japanese PP investor from 1990-2002 how "safe" his/her portfolio was and what the sustainable withdrawal rate would have been to keep from depleting any capital value after inflation (hint...it would be a NEGATIVE SWR as far as I can tell).

I hoped that the Japanese long term treasuries would have saved the day? The yield fell from 7% to 0.7% didn't it? Also deflation would have to be added. So each deflationary year the draw down would have amounted to a smaller nominal amount.

From 1990 to 2000 or so Japan actually averaged slight inflation each year (although less than, say, the US had during the same period). There were some years of deflation and some of inflation but the annual average was slightly inflationary. Now, from circa 2000 to 2012 the average trend was indeed deflationary but those years weren't too bad for the Japanese PP anyway since 2003--mid 2007 were great times for Japanese stocks and since gold was going up sharply during this period as well.

LTTs alone couldn't have "saved the day" in the  1990s for a Japanese PP investor because:

A: The gains in LTTs would be trying to counteract two falling assets (stocks and gold) and one rising asset is not usually enough to counteract two falling assets,

and,

B.  Since LTTs can't fall below whatever rate STTs are at (plus a certain amount or interest percentage to account for the interest rate and duration risk that LTTs have but STTs don't) and STTs cannot really fall below zero for any appreciable length of time, then LTTs can never fall far enough in a deflationary deleveraging environment to offset the other two volatile assets falling sharply. Now if there was no such thing as physical cash, then STTs could indeed fall to, say, -6%, LTTs could fall to -3% (they would pay a less negative rate than STTs since there is a premium for duration and interest rate risk), and deflation might be at -2 or -3%. But since there IS such a thing as physical cash, STT and LTT rates can only fall so far (no matter how bad deflation gets), because if they went very negative for very long then everyone would just hold paper cash instead. Welcome to the zero lower bound.

[stone]

D1984, are you really sure about the LTT not being able to save the day? Gold saved the UK PP in the 1970s when both bonds and stocks were in free fall. The fall in Japanese long term treasury yields from 7% to 0.7% in 1991 to 2001 is the most spectacular bond market rally imaginable isn't it?
If 10year JGB were the longest duration available for retail savers at that time, then the Japanese PP could be modeled as 50% 10year JGB instead of 25% 30year and 25% cash.

[WiseOne]

Thanks for the pointer to the Excel "if" statement!

Did you subtract the cash withdrawal before rebalancing, or take 1/4 from each asset?  If the former, that is a spectacular demonstration of the benefits of using rebalancing bands.  Note that with many Boglehead style "slice and dice" portfolios, annual rebalancing ends up being a necessity.  It's unlikely that adding bits of this and that increases the returns enough to make up the difference.

Before people run out and start assuming a 5% withdrawal rate is safe...it would be worthwhile to incorporate expenses into the simulation - including fund expense ratios, buy/sell fees, gold storage costs....and TAXES.  When I go to calculate total savings, I discount tax-deferred accounts by 20%, for starters.

[iwealth]

Is there a reason 1972-1974 is included in simulations of portfolios that hold a significant amount of gold? If I understand correctly, it was illegal to own gold before January 1, 1975. Doesn't that fudge the numbers significantly in favor of the PP those first few years?

[Tyler]

Did you subtract the cash withdrawal before rebalancing, or take 1/4 from each asset?  If the former, that is a spectacular demonstration of the benefits of using rebalancing bands.  Note that with many Boglehead style "slice and dice" portfolios, annual rebalancing ends up being a necessity.  It's unlikely that adding bits of this and that increases the returns enough to make up the difference.

Before people run out and start assuming a 5% withdrawal rate is safe...it would be worthwhile to incorporate expenses into the simulation - including fund expense ratios, buy/sell fees, gold storage costs....and TAXES.  When I go to calculate total savings, I discount tax-deferred accounts by 20%, for starters.

Yes, for the withdrawal rate analysis I always took expenses out of cash.  Think of it as withdrawing your expenses for the next year on Dec31st and rebalancing the next day (if necessary).  For the diverging PP plot above, I did the "usual" retirement calculation with the blue line -- calculating the total annual gains based solely on the top-line percent gain for the portfolio as a whole (which assumes a clean rebalance every year) and subtracting expenses from the total.

And I WHOLEHEARTEDLY agree with your note about other taxes and fees.  Everyone must evaluate their own personal situation and adjust accordingly.  Taxes are just so complex that it's impossible to make a model that works the same for everyone.  On the positive side, though, the PP is still more tax-efficient than most portfolios so at least you're on the right track.  And smart tax planning during rebalance events (like splitting it over Dec/Jan to minimize the gains in any one year) can make a big difference.  As I've said a few times, it's all about staying smart and flexible.

[Tyler]

Is there a reason 1972-1974 is included in simulations of portfolios that hold a significant amount of gold? If I understand correctly, it was illegal to own gold before January 1, 1975. Doesn't that fudge the numbers significantly in favor of the PP those first few years?

The US came off the gold standard in 1971, but I agree that the resulting wild gold years after that may not ever be repeated and thus may be deceiving.  Likewise, I hate it when stock bugs start any investment examples in the 80's and 90's.  That's why when discussing things like this I prefer to look at scenarios where you retired in a variety of years and economic conditions. 

[HB Reader]

Is there a reason 1972-1974 is included in simulations of portfolios that hold a significant amount of gold? If I understand correctly, it was illegal to own gold before January 1, 1975. Doesn't that fudge the numbers significantly in favor of the PP those first few years?

Owning gold bullion was prohibited prior to January 1, 1975, but it had been perfectly legal since the original prohibition in the 1930's to own certain gold coins, provided they were struck before certain dates.  These exceptions were built into the law primarily to avoid disrupting numismatic activities, as well as for the administrative (enforcement) convenience of the Treasury Department.

Many of these, like the British Sovereign (~.23 oz of gold) and the Mexican 50 peso (~1.2 oz of gold) always traded very close to their bullion value since they weren't particularly rare.  I bought some of both coins prior to 1975 at very close to their bullion value with no problems whatsoever.   

Very few Americans bought legal gold coins until the early 1970s, primarily because it was the functional equivalent of keeping your cash in non-interest bearing status since the London gold price was being held at about $35 (and later, for a short period, at about $42) an ounce under the old Bretton Woods monetary arrangements.  The Bretton Woods system disintegrated fairly gradually in stages in the early 1970s.

As HB pointed out (in a slightly different context) in How I Found Freedom in an Unfree World, there are often legal ways to accomplish things you want to do despite superficial legal or social restrictions, if you are willing to look for them.           

     

[One day at a time]

PS,

Great post.  No doubt, your math adds up for the vast majority.  I wanted to add Bill Bernstein's book to this argument:

http://www.amazon.com/The-Ages-Investor ... +bernstein

I think Bernstein is trying out a digital only distribution for short, chapbook kind of content covering a single topic for more serious investors.  When I read that book, it made a big (and depressing) impact on our family's investing outlook.  Actually, our situation was one of the case examples in the "good luck" camp for ever being able to retire.

A couple of random thoughts/conclusions we made:

1.  I plan to work past retirement age as MT and others have suggested.  It seems that is the easiest way to make math, ie. compounding, work for you.  With that said, the percentage of people disabled enough to preclude  work by 65 is quite high...I think around 1/3.  Even if one lives a healthy life, this does not rule out disability.  Also, the big one here that people overlook is that they may very well be caregivers for a disabled family member.
2.  We decided to have a series of "rip cords" in place to use depending on where we are at retirement age with work and health.  So, here's what we have done (I am 42):

-Currently developing job skill that I could do at any age/disability level as long as I was not cognitively impaired.
-Buying an older annuity, ie. kicks in at 85, in case the money runs out.
-Being clear with our kids (and our own intentions) that the greatest gifts we can give our kids is our time now and avoiding being a financial burden when they are holder.  As someone who works in palliative care, I can say this last point with some experience...avoiding being financially dependent on your children is a much better gift than an inheritance.  Obviously. with the idea that we will spend down our portfolio, that allows for a higher withdrawal rate.  We'll probably annuitize part of it as well.
-Possibly move to a really cheap, retirement friendly country for a period of time to stretch the portfolio.
-Understanding that we will likely sell our house for cash at some point.  I see so many sick patients paralyze their family's ability to help them by their unwillingness to sell their house for liquidity. 

Anyway, those are just some ideas we've focused on.  Looking forward to what better ideas people have considered.