Golden Butterfly Portfolio

A place to talk about speculative investing ideas for the optional Variable Portfolio

Moderator: Global Moderator

barrett
Executive Member
Executive Member
Posts: 1982
Joined: Sat Jan 04, 2014 2:54 pm

Re: Golden Butterfly Portfolio

Post by barrett » Sun Aug 28, 2016 7:49 am

Ocho, Excellent work! Did you by chance run a 100% stocks scenario? That is often argued for (sometimes with passion) over on the MMM forum. Didn't see it on your list but maybe I just missed it.

Also - and this question is for anyone who wants to answer - I haven't really hovered up everything written on the GB, but isn't much of its attractiveness due to the amazing performance of stocks in general (forget about SCV, LCB, etc. for a moment) from 1982 to 2000?
User avatar
ochotona
Executive Member
Executive Member
Posts: 3353
Joined: Fri Jan 30, 2015 5:54 am

Re: Golden Butterfly Portfolio

Post by ochotona » Sun Aug 28, 2016 8:15 am

100% stocks did poorly.

I added a variant of the M. Faber Ivy portfolio, which would really make your heirs happy, or protect you from huge end-of-life health care cost escalation, yet give you a strong initial withdrawal rate.

30% mid-cap blend
30% international small cap
10% REIT
10% gold
20% aggregate bond

Image
barrett
Executive Member
Executive Member
Posts: 1982
Joined: Sat Jan 04, 2014 2:54 pm

Re: Golden Butterfly Portfolio

Post by barrett » Sun Aug 28, 2016 8:28 am

ochotona wrote:100% stocks did poorly.
Got it and thanks for posting that. Interesting that the "Income" portfolio has done so well. Going forward that would not seem possible based on where rates are now. I guess the old disclaimer applies here... Backtesting can tell us what has not worked well but it can't tell us what will work well in the future. Still holding my PP cards for now.
Kbg
Executive Member
Executive Member
Posts: 2815
Joined: Fri May 23, 2014 4:18 pm

Re: Golden Butterfly Portfolio

Post by Kbg » Sun Aug 28, 2016 8:46 am

barrett wrote:Also - and this question is for anyone who wants to answer - I haven't really hovered up everything written on the GB, but isn't much of its attractiveness due to the amazing performance of stocks in general (forget about SCV, LCB, etc. for a moment) from 1982 to 2000?
I've written about this before, but with regard to SCV it is somewhat like gold in the very early 1970s. Once something becomes exploitable its ability to generate returns decreases, sometimes significantly. I'm a fan of rolling 10 year average returns as a way to analyze a portfolio. If a port's 10 year rolling returns stay stable then you probably have something that continues to work. If they do nothing but decrease then there is cause for concern. I did an analysis of a standard HBPP through 2015 and interestingly it's rolling 10 year is increasing after bottoming out in 2001. That's an extremely good sign and somewhat rare.


According to my calcs the latest 10 year rolling return for the PP is 6.98 with a low in 2001 of 6.47. The max occurred in 1982 at 13.29%. On the downside last year was the worst 3 year rolling average ever at 1.39%. If we come in close to where we are now then the rolling 3 year will go back to normal in the 6ish range.

Update/Edit

The latest 3/5/10 year rolling returns are (through 2015):

GB: 3.97/5.74/7.36%
PP: 1.39/4.29/6.98%

Comparing 100% SCV to Large Cap Blend and Total Market on the PV website has SCV with lowest 5 year rolling return by a little under 2% points.

Is it Tyler who builds that website? If so, could you add a 10 year rolling to the 3 and 5?

My personal opinion is the big kicker is 40% stocks more than anything vs. a PP. But I digress.

The main point of this analysis was the value added of SCV. If you analyze the rolling returns of SCV they fall off the proverbial cliff in 1988 as compared to their previous history going back to 1972. Anyone want to guess what happened at the end of June 1987? The advent of the Russell 2000 indexes. Accordingly, a more accurate use/estimate of returns using historical data for SCV should probably start in 1988. Checking average returns since then as well as a 20 and 10 year look back we find:

SCV: 13.06/11.45/9.17
LCB: 11.68/9.83/9.01
TMkt: 11.78/9.97/9.30

And that's what the data has to say on the issue...
User avatar
Tyler
Executive Member
Executive Member
Posts: 2066
Joined: Sat Nov 12, 2011 3:23 pm
Contact:

Re: Golden Butterfly Portfolio

Post by Tyler » Sun Aug 28, 2016 11:21 am

Kbg wrote: Is it Tyler who builds that website? If so, could you add a 10 year rolling to the 3 and 5?
I run Portfolio Charts. Portfolio Visualizer is also a nice tool but I have no affiliation with it.

Try this Rolling Returns calculator, and just set the timeframe to 10. I agree that the trends are very insightful. You'll probably also appreciate the Start Date Sensitivity calculator that is also based on rolling 10-year timeframes.
User avatar
ochotona
Executive Member
Executive Member
Posts: 3353
Joined: Fri Jan 30, 2015 5:54 am

Re: Golden Butterfly Portfolio

Post by ochotona » Sun Aug 28, 2016 1:24 pm

Tyler, your Withdrawal Rates calculator doesn't do a Monte Carlo simulation, correct? It looks at different combos of actual 10-40 year retirements?

That would explain why portfoliovisualizer's withdrawal rates are more pessimistic than your portfoliocharts rates... because by doing the actual simulation the PV site is filling hypothetical tail probabilities that were not observed from the 1970s until the present day.

What's lacking in both of these sites is an ability to do Safe Withdrawal Rate modeling on a trend-following portfolio. The author of portfoliovisualizer is coming to Houston in a few days, I'm going to his talk at the Houston AAII, and I'm going to ask for this feature.

My intuition tells me that trend-following will move any portfolio to the right and maybe also upward on the chart I posted. Less drawdown and volatility will make it less likely that the portfolio will bottom out and "game over".

Absolute Momentum I think will move portfolios to the right. Dual Momentum to the right and up, I believe, but it's going to end up being a tug-of-war between the inherent volatility of the Dual Momentum approach (significant) and how much improved the return is (also significant).

These unknowns make it really hard for the retiree to budget. One of my concerns is that I over-save for retirement, and work for years longer than necessary, because I can't properly assess my trend-following portfolio.
User avatar
Tyler
Executive Member
Executive Member
Posts: 2066
Joined: Sat Nov 12, 2011 3:23 pm
Contact:

Re: Golden Butterfly Portfolio

Post by Tyler » Sun Aug 28, 2016 2:42 pm

ochotona wrote:Tyler, your Withdrawal Rates calculator doesn't do a Monte Carlo simulation, correct? It looks at different combos of actual 10-40 year retirements?

That would explain why portfoliovisualizer's withdrawal rates are more pessimistic than your portfoliocharts rates... because by doing the actual simulation the PV site is filling hypothetical tail probabilities that were not observed from the 1970s until the present day.
Yes, I use actual historical returns and sequence of returns. The default Portfolio Visualizer Monte Carlo simulation method uses the same pool of actual historical returns but randomizes them. So it does not model unobserved fat-tail returns -- it only mixes known returns up in unobserved ways.

Personally, I don't care for that method because I'm not convinced it accurately models the ebb and flow of a real economy. I'm also intuitively concerned that a sample size of only 44 is insufficient for a meaningful collection of 10k different randomized runs, as one bubble or crash will be permanently over-represented. I prefer to offer actual historical results and provide necessary context rather than veer too much into hypothetical models not grounded in reality. But I do appreciate having multiple methods at our disposal to look at the same problem.

Adding various rebalancing and trend following options is on my radar, but I'm still thinking through how to make them most useful for people without making things unnecessarily complex. For me the message is more important than the tool.
User avatar
MachineGhost
Executive Member
Executive Member
Posts: 10054
Joined: Sat Nov 12, 2011 9:31 am

Re: GB wins the portfoliovisualizer.com SWR shoot-out

Post by MachineGhost » Sun Aug 28, 2016 3:40 pm

Hmm, more terminal wealth but lower SWR or higher SWR and lower terminal wealth.

The PP sure sucks.

How about finding the sweet spot in terms of asset weights where the SWR and terminal wealth are both maximized? That would be the optimal portfolio. Tentatively, that looks like the FundAdvice portfolio. What is it doing differently?
"All generous minds have a horror of what are commonly called 'Facts'. They are the brute beasts of the intellectual domain." -- Thomas Hobbes

Disclaimer: I am not a broker, dealer, investment advisor, physician, theologian or prophet.  I should not be considered as legally permitted to render such advice!
User avatar
MachineGhost
Executive Member
Executive Member
Posts: 10054
Joined: Sat Nov 12, 2011 9:31 am

Re: Golden Butterfly Portfolio

Post by MachineGhost » Sun Aug 28, 2016 3:46 pm

ochotona wrote:I added a variant of the M. Faber Ivy portfolio, which would really make your heirs happy, or protect you from huge end-of-life health care cost escalation, yet give you a strong initial withdrawal rate.

30% mid-cap blend
30% international small cap
10% REIT
10% gold
20% aggregate bond
Woah, Nelly! Is this with the market timing?
"All generous minds have a horror of what are commonly called 'Facts'. They are the brute beasts of the intellectual domain." -- Thomas Hobbes

Disclaimer: I am not a broker, dealer, investment advisor, physician, theologian or prophet.  I should not be considered as legally permitted to render such advice!
User avatar
MachineGhost
Executive Member
Executive Member
Posts: 10054
Joined: Sat Nov 12, 2011 9:31 am

Re: Golden Butterfly Portfolio

Post by MachineGhost » Sun Aug 28, 2016 3:52 pm

ochotona wrote:My intuition tells me that trend-following will move any portfolio to the right and maybe also upward on the chart I posted. Less drawdown and volatility will make it less likely that the portfolio will bottom out and "game over".

Absolute Momentum I think will move portfolios to the right. Dual Momentum to the right and up, I believe, but it's going to end up being a tug-of-war between the inherent volatility of the Dual Momentum approach (significant) and how much improved the return is (also significant).
Market timing normally moves risk away from the left fat tail, but higher up is always questionable. Dual Momentum is the same fundamental calculation just done differently so it has no advantage in moving returns upwards. From what I remember, IVY5 benefits from including REIT as a separate asset class in hindsight which was responsible for nearly all of its huge outperformance.

It just keeps coming back to... since we can't forecast the future, you need to be equity diversified. But if you get the timing wrong for equity diversification such as since 2009, its going to kill your net returns and that is dangerous to SWR and terminal wealth. Whether or not market timing can help with that problem is also questionable -- I have seen it go both ways.

In all cases I've seen, adding market timing to the PP doesn't do much of anything other than to reduce the risk. The real advantage comes in applying it to targeted investing.

That being said, I do use market timing on the PP because I think leaving money on the table is bloody stupid (which is what you do when you give back your equity gains while relying on T-Bonds and/or gold to overcome those losses). Since its not going to go to other investors as they don't "believe" in market timing, I sure as hell am not going to give it to Wall Street if I can help it. However, its got to be worth the time and expense of extra effort and slightly lowered returns, so I've amped it up in the Prosperity sector as I've noted several times. My latest consideration for my core level is tripling the S&P 500 returns with 50% less maximum drawdown which is finally "good enough" (and thank gawd, for I was getting stressed...)

Anyway, what was the core outperformer in your modified IVY5 since REIT was only 10%?
"All generous minds have a horror of what are commonly called 'Facts'. They are the brute beasts of the intellectual domain." -- Thomas Hobbes

Disclaimer: I am not a broker, dealer, investment advisor, physician, theologian or prophet.  I should not be considered as legally permitted to render such advice!
User avatar
ochotona
Executive Member
Executive Member
Posts: 3353
Joined: Fri Jan 30, 2015 5:54 am

Re: Golden Butterfly Portfolio

Post by ochotona » Sun Aug 28, 2016 5:56 pm

MachineGhost wrote:
ochotona wrote:I added a variant of the M. Faber Ivy portfolio, which would really make your heirs happy, or protect you from huge end-of-life health care cost escalation, yet give you a strong initial withdrawal rate.

30% mid-cap blend
30% international small cap
10% REIT
10% gold
20% aggregate bond
Woah, Nelly! Is this with the market timing?
No market timing. The source of out-performance is both (1) US mid-cap blend (value is even better, but I just chose blend because portfolio visualizer has a longer history), and (2) small-cap international. We love the mid-caps.
Kbg
Executive Member
Executive Member
Posts: 2815
Joined: Fri May 23, 2014 4:18 pm

Re: Golden Butterfly Portfolio

Post by Kbg » Mon Aug 29, 2016 9:54 am

I'm not going to lay out the homework this time but if you think international small caps are going outperform relatively as they have historically you are going to be disappointed. The reason is the same as SCV..this sector has been commoditized.

A better line of study if you feel you have the ability to add some value from some type of timing mechanism is to look at what drives small cap vs large cap performance domestically and what drives domestic vs. foreign performance. Another fruitful area is sector research. There is huge overlap between foreign/domestic and sector performance. For example, when commodities do well certain countries do well...but there are a lot of other relationships if one digs.

However, if you are up for just a little more work it is easy to outperform indexes with some very simple stock screens or just buy an equal weight ETF. Most people have a hard time with tracking error though.
User avatar
ochotona
Executive Member
Executive Member
Posts: 3353
Joined: Fri Jan 30, 2015 5:54 am

Re: Golden Butterfly Portfolio

Post by ochotona » Mon Aug 29, 2016 10:33 am

Kbg wrote:I'm not going to lay out the homework this time but if you think international small caps are going outperform relatively as they have historically you are going to be disappointed. The reason is the same as SCV..this sector has been commoditized.

A better line of study if you feel you have the ability to add some value from some type of timing mechanism is to look at what drives small cap vs large cap performance domestically and what drives domestic vs. foreign performance. Another fruitful area is sector research. There is huge overlap between foreign/domestic and sector performance. For example, when commodities do well certain countries do well...but there are a lot of other relationships if one digs.

However, if you are up for just a little more work it is easy to outperform indexes with some very simple stock screens or just buy an equal weight ETF. Most people have a hard time with tracking error though.
That just untrue. SCHC (intl small cap) has been beating SCHF (intl developed market) over the past year. As a momentum person, that's evidence I pay attention to.
User avatar
MachineGhost
Executive Member
Executive Member
Posts: 10054
Joined: Sat Nov 12, 2011 9:31 am

Re: Golden Butterfly Portfolio

Post by MachineGhost » Mon Aug 29, 2016 11:43 am

Kbg wrote:A better line of study if you feel you have the ability to add some value from some type of timing mechanism is to look at what drives small cap vs large cap performance domestically and what drives domestic vs. foreign performance. Another fruitful area is sector research. There is huge overlap between foreign/domestic and sector performance. For example, when commodities do well certain countries do well...but there are a lot of other relationships if one digs.
So global sectors would be superior to domestic? They're market cap weighted though. I wonder about the currency exposure though.

What about using equal weight domestic sectors instead of market-cap weight?
"All generous minds have a horror of what are commonly called 'Facts'. They are the brute beasts of the intellectual domain." -- Thomas Hobbes

Disclaimer: I am not a broker, dealer, investment advisor, physician, theologian or prophet.  I should not be considered as legally permitted to render such advice!
Kbg
Executive Member
Executive Member
Posts: 2815
Joined: Fri May 23, 2014 4:18 pm

Re: Golden Butterfly Portfolio

Post by Kbg » Mon Aug 29, 2016 12:13 pm

MG,

As you probably know international and US stock market correlation is extremely high now as is US/Intl sector correlation. My main point was that if you compare a country index to a sector index the country index is often highly correlated to a specific sector. For example, CAN and AUS are highly correlated to energy/commodities. The US is highly correlated to tech and the UK is highly correlated to finance. So if you are investing internationally and a specific sector is on a tear then it may be useful to look for a highly correlated country.

With regard to equal vs. cap weight. On broad indexes there is no doubt equal weighting is better even after costs which are higher with EWs. This effect is more pronounced in large cap indexes than their smaller brethren for obvious reasons.

Sectors aren't as straight forward, but the general rule applies unless a sector is completely dominated by a very competitive company(s). Think Apple in tech. However, I'm willing to put a milk shake on the line that this won't hold over the long haul because as soon as that dominant company begins to enter maturity return to mean is going to happen. Of course, a domination run can last a very long time.

There is some pretty good analysis on US cap vs. EW sectors over on Seeking Alpha. http://seekingalpha.com/article/4002723 ... onclusions. The stats and analysis at the end are quite good.

Finally,

No one but no one knows what the future brings and while backtests are interesting they are "back" tests. If someone is interested in performance and can accept a bit more volatility I think there are some very simple ways to beat future fixed portfolio averages with different levels of aggressiveness driving how management of the portfolio would be conducted.

One could start out with a basic strategic asset allocation decision that would dedicate a certain range of the portfolio to a given asset class or asset classes. Volatility of the portfolio would be calibrated by the amount of the port dedicated to cash or home country STTs. Using PP as a template (and there are others that could be used). One could divide the non-cash PP into stocks and what I will call cash substitutes (gold and LTTs). Once the strategic framework has been put into place then specific assets are selected for the various buckets. Keeping PP as our strategic framework we will place GLD and TLT in one bucket and EEM, SPY and IWM in another bucket. After our assets are chosen we will then look for ways to get a ride on very long term secular trends. A good way to do this is to make ratios of all of our assets within their buckets. So in this example GLD/TLT and SPY/EEM and SPY/IWM and EEM/IWM. Then we will monitor these ratios for signs of relative outperformance over a long period of time. We are setting this up for years not weeks or month. Some examples of how this could be done is using a ratio long term moving average or a ratio breakout method like bollinger bands or Donchian channels. So if our GLD/TLT ratio crossed above a 500 day moving average of the same then we would flip to gold, a cross below means a flip to TLT. Our SPY ratios will tell us if foreign or small caps are outperforming and if both are then there is the EEM/IWM tie breaker ratio. We can get more cosmic than the above binary approach and set up ranges where we might be 100% in an asset or some percentage of the two. We could get really cosmic and drive our whole portfolio this way or just do range rebalancing between our asset class picks/weights.

So what does the above do? Well we have no idea if it will outperform the hottest asset of whatever future time period we are investing in. In fact, the odds are highly unlikely that will be the case. However, as compared to a fixed approach we will certainly do better than if we picked the wrong asset class. In sum, we have set up a process whereby we are going to benefit from the majority of whatever Mr. Market is going to give us even though we have no idea what this might be in the future or what asset classes will be providing it and when.

One note of caution. There is great temptation to throw in more assets in hopes of riding the best asset. This creates a lot more work, but also dilutes the impact of what we are trying to accomplish and drives up trading frequency significantly. This approach does not chase the hottest asset, it rides major long term secular trends. So assets should be few in number and represent major market components. If you throw too much in a bucket you end up catching the wrong side of many smaller return to mean trades.
User avatar
buddtholomew
Executive Member
Executive Member
Posts: 2464
Joined: Fri May 21, 2010 4:16 pm

Re: Golden Butterfly Portfolio

Post by buddtholomew » Mon Aug 29, 2016 1:22 pm

I just don't see how different this approach is to establishing an asset allocation, selecting the appropriate investment vehicles and re-balancing when bands are triggered.
Kbg
Executive Member
Executive Member
Posts: 2815
Joined: Fri May 23, 2014 4:18 pm

Re: Golden Butterfly Portfolio

Post by Kbg » Mon Aug 29, 2016 1:39 pm

You are going all in to whatever is leading in your bucket...trend following. The alternative is look at history and hope it repeats when picking your asset classes and weights.
User avatar
frugal
Executive Member
Executive Member
Posts: 947
Joined: Sat Nov 10, 2012 12:49 pm

Re: Golden Butterfly Portfolio

Post by frugal » Thu Sep 01, 2016 4:20 am

Hello my friends!

How are you?

Any European has changed to Golden Portfolio?

It seems interesting...

Regards!

:)
bedraggled
Executive Member
Executive Member
Posts: 705
Joined: Sat Sep 13, 2014 4:20 am

Re: Golden Butterfly Portfolio

Post by bedraggled » Sun Nov 13, 2016 6:00 pm

How is the GB doing since the election?

And compared to the 4x25 HBPP?
User avatar
ochotona
Executive Member
Executive Member
Posts: 3353
Joined: Fri Jan 30, 2015 5:54 am

Re: Golden Butterfly Portfolio

Post by ochotona » Sun Nov 13, 2016 7:16 pm

GB is +0.3%, PP is -0.25% since Tuesday... but that's an insignificant amount of time. It means nothing. Ignore it.
bedraggled
Executive Member
Executive Member
Posts: 705
Joined: Sat Sep 13, 2014 4:20 am

Re: Golden Butterfly Portfolio

Post by bedraggled » Sun Nov 13, 2016 7:46 pm

Thanks, Ocho.
User avatar
buddtholomew
Executive Member
Executive Member
Posts: 2464
Joined: Fri May 21, 2010 4:16 pm

Re: Golden Butterfly Portfolio

Post by buddtholomew » Sun Dec 04, 2016 1:24 pm

A strong USD has had a negative impact on the PP following the elections. Some assets have benefited from this advance, none more so than small caps.

I have rebalanced out of small caps in retirement accounts and am considering adding to taxable. I would prefer to add international stocks, but small cap appears to assist the PP in times of a rising currency whereas international would respond in similar fashion to gold. Obviously these relationships are not concrete.

So, basically 2 questions.
Does the strong USD/small cap relationship hold true and is holding international (unhedged) for a US investor redundant for the PP?
User avatar
Tyler
Executive Member
Executive Member
Posts: 2066
Joined: Sat Nov 12, 2011 3:23 pm
Contact:

Re: Golden Butterfly Portfolio

Post by Tyler » Mon Dec 05, 2016 9:46 am

buddtholomew wrote: So, basically 2 questions.
Does the strong USD/small cap relationship hold true and is holding international (unhedged) for a US investor redundant for the PP?
I don't really have an opinion on question #1. For #2, I've personally come to a similar conclusion that you have. Gold is negatively correlated to US stocks but more positively correlated to international stocks, and I believe that's a structural thing and not simply chance. So in the context of the PP or GB I find adding international doesn't really move the needle in a noticeable way. But for non-PP investors with a strong aversion to gold, I think international stocks are a good option for diversifying away from US stocks.
User avatar
buddtholomew
Executive Member
Executive Member
Posts: 2464
Joined: Fri May 21, 2010 4:16 pm

Re: Golden Butterfly Portfolio

Post by buddtholomew » Mon Dec 05, 2016 2:24 pm

Thanks Tyler.
I am trying to identify an asset that correlates well with a rising dollar and so far only UUP ETF appears to meet the requirement.
Jack Jones
Executive Member
Executive Member
Posts: 522
Joined: Mon Aug 24, 2015 3:12 pm

Re: Golden Butterfly Portfolio

Post by Jack Jones » Mon Dec 05, 2016 3:13 pm

Tyler wrote:
dualstow wrote: I'm out of date. I thought you were just running a pp.
I still am! I just added a small cap VP a few months ago. ;)
Seriously though, as long as you keep your stock allocation at 25%, I feel like you're free to implement the stock portion as you see fit, and I'd still consider it a PP. Even Harry went through some iterations of how the stock portion should be allocated before he settled on just putting the full 25% in the SP500 (perhaps just for simplicity's sake?).

Edit: Whoops, it appears that I am living in the past.
Post Reply