Why 25-25-25-25 weightings?

[Clive]

...

[craigr]

What were the odds for a Japanese investor that their stock market would win each year the past 80 years?

I understand the point though. I suppose part of it from my own analysis is that the weightings provide a smooth ride and simplicity and for many investors that's a key point. I don't deny that when stocks are working they can really pile on returns, but during a protracted dry spell they are very demoralizing.

I suppose the better question to ask is over an investor's saving horizon of 20-40 years, how likely is it that it could include a protracted period of zero real growth for stocks affecting their overall savings? Or what would the effect of a large drawdown be during retirement when they aren't able to save? I'd have to spreadsheet it all, but I know over the last 40 years we've already had two decades of the 1970s and 2000s in the US with bad real returns. So that right there tells me the idea of stocks are the most likely to outperform idea is on thin ice.

[Pointedstick]

For me, the lack of volatility is as big a feature as the good average return. I agree with craigr about stocks really delivering when they're hot, but a 45% stock allocation also opens you up to bigger potential losses during protracted slumps. If I were looking to juice the returns of a PP (and I'm not), I'd replace the 25% TSM allocation with 13% emerging markets and 12% small-cap value stocks. Over the past 40 years, that allocation has actually beaten an orthodox PP in both returns and volatility! But of course, the future is unknowable, which is why I stick with a 25/25/25/25 allocation.

[craigr]

What were the odds for a Japanese investor that their stock market would win each year the past 80 years?

From a quick scan for Japan 1972 to 2008, somewhat surprisingly, stocks were the best performing asset in 48.6% of years (27% of years had gold as the winner and just 2.7% of years for cash).

Weighting to those allocations also seems to have produced reasonable overall rewards of around 3.6% real annualised over that span of years and reasonably linear real growth over the total range, excepting two up/back-down spikes/dips (1986 to 1989 unusually large up spike, 1990 large down dip; 2004 to 2007 unusually large up spike, 2008 large down dip).

How'd it do through WWII?

It's interesting because we use this data largely from U.S. markets, but even then it's kind of thin. However if you start looking at other countries the stock markets are not so robust. A lot of very bad things have happened along the way to make stock investing extremely uncertain in most places of the world!

[dualstow]

Well, Clive, you now know how to fill your VP. 

[Gosso]

For me, the lack of volatility is as big a feature as the good average return

Simba's spreadsheet data for a yearly rebalanced asset allocation of
15% LCV, 15% SCV, 15% EM, 20% LTT, 20% Gold, 15% 2 year T

                                  CAGR   Std.Dev Sharpe Ratio
Portfolio Growth - Nominal 12.18  9.74    0.74
Portfolio Growth - Real    7.51   9.26   0.75
Coffee House - Nominal   10.43  11.27   0.50
Coffee House - Real       5.83   11.32    0.48

Slice 'n' Dicer!  For shame! 

If you simply go with, 45% TSM, 20% LTT, 15% 2 year T, 20% gold, you get (between 1972-2010):

                                    CAGR     Std.Dev Sharpe Ratio
Portfolio Growth - Nominal 10.10%  9.33      0.53

The 4x25% PP gives the following:

                                    CAGR     Std.Dev Sharpe Ratio
Portfolio Growth - Nominal 9.66%  8.10      0.54

I'm a little surprised there isn't a bigger difference, since prosperity has mostly dominated between 1972-2010.

[ngcpa]

I suspect you would do much better with 0 cash and 1/3 each the other 3, although your volatility would be a bit higher.  My backtesting (for about the past 20 years) shows the higher volatilty would be well worth it.

[AdamA]

I suspect you would do much better with 0 cash and 1/3 each the other 3, although your volatility would be a bit higher.  My backtesting (for about the past 20 years) shows the higher volatilty would be well worth it.

Right, but wouldn't you still have some cash saved up somewhere? 

The nice thing about the PP is that you can achieve a decent rate of return even with some emergency cashed figured in.

[Alanw]

I suspect you would do much better with 0 cash and 1/3 each the other 3, although your volatility would be a bit higher.  My backtesting (for about the past 20 years) shows the higher volatilty would be well worth it.

Right, but wouldn't you still have some cash saved up somewhere? 

The nice thing about the PP is that you can achieve a decent rate of return even with some emergency cashed figured in.

Any cash that I have whether it is in IRA, CD, checking account, etc., I consider it part of my 25% PP allocation.  Being semi-retired, I draw on the cash portion for some living expenses.  So far, no need to rebalance.

[EdwardjK]

If equities have delivered superior results over the longterm compared to the three other assets, then tilting the weightings towards equities is not unreasonable.  Just be aware that a bad year in equities will drive down your results compared to the traditional PP formula.

At the same time, why is it not reasonable to use leverage within the PP?  There are ETFs that use leverage to achieve 2x gold, and 2x LT Treasuries, etc..

If 1x the PP is smart, why should we use leverage to achieve 2x PP?

[Pointedstick]

If equities have delivered superior results over the longterm compared to the three other assets, then tilting the weightings towards equities is not unreasonable.  Just be aware that a bad year in equities will drive down your results compared to the traditional PP formula.

At the same time, why is it not reasonable to use leverage within the PP?  There are ETFs that use leverage to achieve 2x gold, and 2x LT Treasuries, etc..

If 1x the PP is smart, why should we use leverage to achieve 2x PP?

You can do whatever you want, but for me, the PP is not only about delivering good results; it's also about limiting volatility and distributing risks. I know that I personally don't react very well to volatile investments; I tend to sell at the wrong times, or doubt my plans and change my asset allocation at the wrong times. a PP composed of 2X ETFs might produce double the returns (but maybe not, due to leveraged ETF decay issues), but it will also double your volatility. Additionally, a leveraged ETF PP will be very vulnerable to brokerage house shenanigans and will obviously prevent you from having any physical gold. You'd have a kinda-sorta-PP, but one with a lot more risk than a an orthodox PP. I might consider jiggering with the PP formula to juice returns for a VP, but not for my core PPs themselves.

[Greg]

I might consider jiggering with the PP formula to juice returns for a VP, but not for my core PPs themselves.

Whenever I feel like my portfolio is feeling a little bit down, I grab for a bottle of genuine Juice Returns.*

*made from concentrate

[Gosso]

I created a synthetic levered PP based on leveraging the daily movements of SPY, TLT and GLD.  (I downloaded the daily closing prices from yahoo.com).  The following graphs do NOT account for borrowing costs, or other fees.

EDIT:  THESE GRAPHS ARE WRONG.  BORROWING COSTS CREATE SIGNIFICANT DRAG.
[img width=600]http://i49.tinypic.com/263zeyp.jpg[/img]

The number of rebalancing events are as follows:
- 1x: 2 times
- 2x: 9 times
- 3x: 14 times
- 5x: 34 times

Max draw-down from "2008" (between May 20, 2008 and March 9, 2009):
- 1x: -15.0%
- 2x: -29.5%
- 3x: -47.0%
- 5x: -70.6%

From my observations the "decay" from the levered ETF's does not pose a problem in a portfolio which is rebalanced.  However, decay does exist as shown by the non-rebalanced levered SPY:

[img width=600]http://i45.tinypic.com/rc3ibl.jpg[/img]

But, if the levered SPY is contained within the 3x33% US PP and is rebalanced at the 40% bands then we get the following:

[img width=600]http://i46.tinypic.com/97qzip.jpg[/img]

I'm not quite sure what to make of all of this.  I have always thought that decay from the levered ETF's made them useless for the "buy-and-hold" investor, but it appears that assumption was incorrect (see this paper posted by Clive, HERE).  I have tested my model against "real-life" levered ETF's (LETF's) and my model actually under-predicts the real LETF's.

I'd appreciate any criticism of my analysis, especially before I do something I will regret.  Any fellow "number-cruncher" that have looked at using LETF's for the PP?  Are there unseen risks that are not obvious?

Edit: fixed first chart since it was cutoff.

[Greg]

That does seem really amazing for that 5x leveraged portfolio. It always looks great in hindsight but I'm wondering psychology if people could stick to this investing plan when you have a 70% drawdown. That could be your entire retirement nest egg and make you panic a lot. This idea might be great as a juice-boosting VP play.

[Gosso]

That does seem really amazing for that 5x leveraged portfolio. It always looks great in hindsight but I'm wondering psychology if people could stick to this investing plan when you have a 70% drawdown. That could be your entire retirement nest egg and make you panic a lot. This idea might be great as a juice-boosting VP play.

I added the 5x because I thought it would be completely destroyed by "2008", due to the volatility drag, but it rebounded like a champ.  I'm not even sure how one would create a 5x PP...I guess you would need to use options, but that opens the door to many other risks which I am not comfortable accepting.

I would like to emphasis that people read through that paper linked to by Clive.  It makes you wonder if 1x is really all that special, and maybe 2x, or 3x is more appropriate for maximizing returns... especially for a portfolio low in volatility.

Obviously LETF's are only appropriate for people that can handle volatility.  I'm young and stupid, and looking for a way to retire early 

[Pointedstick]

Obviously LETF's are only appropriate for people that can handle volatility.  I'm young and stupid, and looking for a way to retire early 

Your best bet is to save 70-90% of your after-tax income. Live a spartan life for a few years, then be set for life.

[Greg]

Obviously LETF's are only appropriate for people that can handle volatility.  I'm young and stupid, and looking for a way to retire early 

Your best bet is to save 70-90% of your after-tax income. Live a spartan life for a few years, then be set for life.

That's what I'm trying to do. Even though I think it'll make me feel as though I won't grow up for a while, I might live with my rents for the next year or two so I only spend around <$5k of my paycheck for the year (most of that going towards gas for a 73 mile round trip each day for work).

[Gosso]

Clive, I greatly appreciate the reply.  The paper that you linked to is HERE.  It raised some interesting questions, such as is 1x optimal, or is it 2x, or 3x, or more?  They only looked at a single asset class in their research, but if you extrapolate the data into a balanced portfolio, then the benefits might be more impressive when compared to a single asset.

Did you account for the cost of borrowing?

I have not accounted for the cost to borrow (I'm not sure how to account for it, do we simply subtract the average short term treasury yield from the LETF returns).  I'm not sure how the LETF's would handle this expense (I skimmed several LETF's prospectus, but couldn't find any satisfactory answer)...I figured they would increase the leverage slightly to account for this so that they could maintain their mandate of 2x or 3x DAILY movement.  So in reality they might be using 2.1 or 2.2x leverage to overcome the cost to borrow and therefore maintain the 2x daily movement...although I might be wrong on this.  Looking at the returns from SPY and SSO, I don't see a significant divergence from their 2x mandate...

[Gosso]

When I measured over widened and subsets however I believe it indicated that that varied between (if memory serves) -5x to +4x for an overall average of -0.5 i.e. deduct -0.5x short from a 100% long and that's 50% stock, 50% cash.

You lost me at -5x...that seems completely insane to my simple mind.  How would a Japanese PP at 2x perform? (I don't have the data).

So, you claim that the LETF's would lose out on the cost to borrow, but I'm not sure...wouldn't they simply increase the leverage to overcome this.  After-all their mandate is to maintain 2x or 3x the DAILY movement, not to maintain 2x or 3x overall leverage.

[Gosso]

i.e. over that full period the 2x and underlying 1x had provided near enough the same overall reward, but clearly the volatility for the 2x was (obviously) much higher.

Right.  I agree that SSO has been slaughtered over the past decade, but combine that lowly SSO into a balanced portfolio of UBT and UGL, and you have created a formidable foe (at least from my backtesting).  

In practice its much deeper as 2x funds might be using derivatives or a whole range of alternatives. Fundamentally it will all boil down to much the same thing however.

Some time back I ran a test



and found that my synthetic was very close to the actual. I then used that as a basis to create a synthetic that went back further in time and saw

Hmm...I have had a difficult time factoring in the cost to borrow into the LEFT returns.  I suppose it could come directly off the returns, but then the LETF's are not meeting their mandate...wouldn't they increase leverage to overcome this?  Looking at your data it appears they don't...I would be pissed if I was a shareholder...

So, in summary, it only makes sense to leverage the PP when real interest rates are negative...ie we are borrowing (stealing) money from grandma to leverage our investments?

[MachineGhost]

That does seem really amazing for that 5x leveraged portfolio. It always looks great in hindsight but I'm wondering psychology if people could stick to this investing plan when you have a 70% drawdown. That could be your entire retirement nest egg and make you panic a lot. This idea might be great as a juice-boosting VP play.

It wouldn't have a 70% portfolio drawdown which is how you would compare to 1x.  I argue they would all have the equivalent maximum drawdown, so there really is no advantage to using leverage funds.  The illusion of security by holding more cash doesn't reduce risk.

[Gosso]

I did some more tinkering with the numbers.  For some reason I got LIBOR+1.75% as a better fit to match the synthetic SSO and UPRO with SPY.  I'll go with this for now, plus it is more conservative.  I also added a 3% dividend to SPY and TLT of the 1x PP.  This is what it looks like.

[img width=600]http://i50.tinypic.com/fo3lmo.jpg[/img]

The 2x begins to pulls away from the 1x once LIBOR drop below ~2%.

Max drawdown:
1x = -13.5%
2x = -34.5%
3x = -55.0%

[Gosso]

More tinkering:

EDIT:  THIS GRAPH IS WRONG.  I'LL LEAVE IT UP SINCE IT SHOWS THE LEARNING PROCESS I WENT THROUGH.
[img width=600]http://i45.tinypic.com/2i6fdrd.jpg[/img]

The borrowing costs of the 70's and 80's destroy the LETF's.  The more I look at the numbers the more I appreciate the plain old 4x25% PP.

However, LETF's do make some sense as of right now (as a part of a VP), due to low borrowing costs.  Just realize that you are risking twice or triple the downside, which may result in returns not much different than the 1xPP, unless you are an astute market timer.

[clacy]

If/when the PP hits another 15% DD I will likely go into a 1.5 or 2x leveraged version with the intention that you could see a huge run up and you're buying in at a relatively cheap position

[Gosso]

If/when the PP hits another 15% DD I will likely go into a 1.5 or 2x leveraged version with the intention that you could see a huge run up and you're buying in at a relatively cheap position

I agree.  It's just a matter of waiting for the PP to stumble.  Also, I think this is good from a psychological point of view, since once the PP does see a large temporary drawdown (10-15%), then I will see this as an opportunity, rather than freaking out and buying canned food, bullets, and whiskey.  So instead of fear taking over, I will allow greed to control my actions.  I like greed more than fear.