Golden Butterfly Portfolio

[Dieter]

I've been spending some more time looking at the GB and various iterations and am wondering why Desert's variation hasn't gotten more discussion. With the obvious caveat that backtesting only goes so far when I run the numbers comparing the GB with Desert's version the latter comes out well ahead on the risk:return scale. Here's one set of numbers from 1975-2015:

https://www.portfoliovisualizer.com/bac ... 0&Gold2=10

Since it's essentially a gold-enhanced version of one of Larry Swedroe's "No Fat Tails" portfolios I played around with substituting other stock allocations and also removing the gold entirely and didn't do nearly as well. I also substituted the intermediate-term treasuries that make up 60% of his allocation for the bond barbell in the GB and it made no difference to the returns, though obviously from a practical rebalancing and/or retiree perspective having 20% of the allocation in SHY or cash makes for greater ease.

Backtesting aside, I'm looking at the 30% highly-diversified stocks, 60% highest-quality bonds of optimum duration in terms of risk:reward sweet spot and just enough gold for (as Desert puts it) some modest SHTF insurance but not enough to be a hassle to own and wondering why this isn't a significant improvement on the GB, especially in retirement.

Just note that the link isn't GB (https://portfoliocharts.com/portfolio/golden-butterfly/) vs Desert (https://portfoliocharts.com/portfolio/desert-portfolio/), if those links are correct...

-- The GB-like one has Small-Balanced instead of Small Value; total US vs LCB
-- The Desert-like one replaces 30% total market with 10% each of Total Market, Small Cap Value, and EM

(not saying bad, but different than what I expected)

Making the GB-Like one actually GB (switch SCB to SCV) makes the comparison closer:
-- Same CAGR
-- Desert-like has lower volatility (60% bonds; 10% Gold)

But who knows if the next 35 years will be a falling rate environment and how Gold performs vs Treasury's (or EM, SCV, US Total Market, ....)

Below link is the above with actual GB. And I threw in a third portfolio, cause, well, I get sucked in when I go to that site, and was something kinda in the middle of the two....

https://www.portfoliovisualizer.com/bac ... 0&Gold3=12

[Edit: added links to GB / Desert portfolios on Portfolio Charts]
[Edit2: switched Total US to LC Balanced in GB in link I provided; added note re: diff]

[sophie]

I thought Desert's portfolio was mainly designed for people who are afraid to hold a lot of gold, and need something simple to manage? In those respects I think it's brilliant. However - look at the portfolio's performance in the 1970s. 10% gold isn't enough. It still beat the pants off any pure stock/bond allocation in that period, which is good enough for the intended audience.

Harry Browne didn't just arbitrarily select the 25x4 allocation. He ran a lot of simulations and backtests before coming to that conclusion - quite a feat at the time. I've never seen details of the results of the testing, though. Between this and the fact that the "secret" about small cap value funds is out of the bag (totally agree with MG on this point), I'm not planning to switch to Golden Butterfly. I also worry that there's a significant element of recency bias. Stocks are great until they aren't. If you look at the S&P 500's returns since 2000, you won't be impressed. I'm always amused at the "stocks reach record highs!" headlines, which sound great until you realize that you can translate this to "stocks finally make money for the first time in decades".

However, I'm considering IJS (small cap value) in taxable, to reduce dividends. It's a commission-free fund at Fidelity. The expense ratio is a lot to stomach though.

[ozzy]

However, I'm considering IJS (small cap value) in taxable, to reduce dividends. It's a commission-free fund at Fidelity. The expense ratio is a lot to stomach though.

You may want to look at IJR (SmallCap Blend). Also commission-free at Fidelity. Broader diversification, low expense ratio (0.12%), and better performance. I've owned it for years, its great.

[Kevin K.]

Tyler actually replaced Small Cap Value with Small Cap Blend in the version of the GB he said he himself has implemented. Aside from the lengthy thread on GB here there are two other equally epic ones on the Mr. Money Mustache and Early Retirement Extreme forums and I don't recall on which of them he mentioned this, but as I recall it was a minor tweak reflecting the fact that TSM already has some value stocks.

Regarding Desert's portfolio, you're referring to "the" desert portfolio featured on Tyler's site, while I'm referring to the one Desert said (earlier in this very thread) he's using now, which is 10% each TSM, SCV, EM and GLD and 60% IT Treasuries. This allocation outperforms the GB in terms of risk:return over any time period I've run the numbers. Unlike GB which is essentially a PP slightly tilted towards stocks this "new and improved" Desert (Desert Oasis? clearly it needs a name to compete with GB ) is basically Larry Swedroe's No Fat Tails allocation spiked with a bit of gold.

[dualstow]

Unlike GB which is essentially a PP slightly tilted towards stocks this "new and improved" Desert (Desert Oasis? clearly it needs a name to compete with GB )

And a date palm logo.

[barrett]

My only issue with the Desert Oasis is that it doesn't hold any cash. As someone who is looking to retire in the next little while, that's an important consideration. One could argue that the treasuries and stocks are kicking off a lot of interest and dividends, I guess. My recollection is that Desert himself is a decade or so shy of pulling the plug on work. Incidentally, have we lost him due to forum outages? Haven't seen a post from him recently.

[Tyler]

Tyler actually replaced Small Cap Value with Small Cap Blend in the version of the GB he said he himself has implemented.

Yep -- that's true. I went back and forth, and ultimately chose a simple split between VTI and VB for the stock portion of my portfolio. When I found myself splitting hairs, I took that as a sign to stop over-thinking it and defaulted to the more neutral option.

[dualstow]

Tyler actually replaced Small Cap Value with Small Cap Blend in the version of the GB he said he himself has implemented.

Yep -- that's true. I went back and forth, and ultimately chose a simple split between VTI and VB for the stock portion of my portfolio. When I found myself splitting hairs, I took that as a sign to stop over-thinking it and defaulted to the more neutral option.

I'm out of date. I thought you were just running a pp.

[Tyler]

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.

[dualstow]

ah, ok.

[koekebakker]

My only issue with the Desert Oasis is that it doesn't hold any cash. As someone who is looking to retire in the next little while, that's an important consideration. One could argue that the treasuries and stocks are kicking off a lot of interest and dividends, I guess. My recollection is that Desert himself is a decade or so shy of pulling the plug on work. Incidentally, have we lost him due to forum outages? Haven't seen a post from him recently.

You can have a Desert portfolio (or any portfolio) with cash. What you want is a target duration. It doesnt matter what mix of fixed income you use to get that duration. I can't remember if Desert used a 5 or 10 year duration but let's say he wants a 5 year duration for his 60% fixed income he can choose to put 60% in 5-year treasuries and don't have any cash, or he can put 30% in 10-year treasuries and 30% in cash. Or he can put 15-20% in LT's and 40-45% in cash to get a duration of 5 years. The performance will be very similar.

[koekebakker]

Regarding the GB: if something looks too good to be true...
SCV backtests so well because of the very high returns over the last 40 years and because of the low correlation. The holy grail basically.
Both are unlikely to repeat themselves.
It can still make sense to hold SCV but there's no reason to believe it will keep outperforming as it did in the past.

[Kevin K.]

Regarding the GB: if something looks too good to be true...
SCV backtests so well because of the very high returns over the last 40 years and because of the low correlation. The holy grail basically.
Both are unlikely to repeat themselves.
It can still make sense to hold SCV but there's no reason to believe it will keep outperforming as it did in the past.

This isn't true, and neither are the other comments dissing EM, IV, ISCV and other stocks vs. TSM found frequently on these forums. The Fama-French three factor model, Bill Bernstein's work and the ongoing research at firms like DFA are not crackpot theories, they're evidence-based approaches to investing that continue to be refined and revised on an ongoing basis. It's perfectly fine not to slice and dice, or to say that you prefer the simplicity of TSM or (a la John Bogle himself) simply see no need to tilt or own international stocks, but to say that such allocations are purely based on backtesting makes exactly as much sense as saying that the PP is just a goldbug's bunker based on 1970's economy combined with libertarian politics.

There are any number of portfolio's on Tyler's site (The Coward and Merriman's Ultimate come to mind first) that illustrate the value of embracing this kind of complexity over time. For those interested in a crash course cum Cliff's Notes summary of the logic of MPT I think this article on cloning DFA equity allocations using Vanguard funds is a useful point of entry:

http://www.etf.com/publications/journal ... nopaging=1

[koekebakker]

Regarding the GB: if something looks too good to be true...
SCV backtests so well because of the very high returns over the last 40 years and because of the low correlation. The holy grail basically.
Both are unlikely to repeat themselves.
It can still make sense to hold SCV but there's no reason to believe it will keep outperforming as it did in the past.

This isn't true, and neither are the other comments dissing EM, IV, ISCV and other stocks vs. TSM found frequently on these forums. The Fama-French three factor model, Bill Bernstein's work and the ongoing research at firms like DFA are not crackpot theories, they're evidence-based approaches to investing that continue to be refined and revised on an ongoing basis. It's perfectly fine not to slice and dice, or to say that you prefer the simplicity of TSM or (a la John Bogle himself) simply see no need to tilt or own international stocks, but to say that such allocations are purely based on backtesting makes exactly as much sense as saying that the PP is just a goldbug's bunker based on 1970's economy combined with libertarian politics.

I agree. They are evidence based but you can debate about how strong the evidence is.
I even believe in Small/Value having a higher expected return than a cap-weighted index. I also believe however, that the level of out-performance of small/value in all those backtest-tools is not likely to be repeated. The difference in CAGR between TSM and SCV is almost 3% a year from 1972-2015! I think it's fair to warn people not to get too excited by all those nice backtested graphs...
Bernstein himself expects the value premium to be much, much lower going forward.

[Kevin K.]

I agree. They are evidence based but you can debate about how strong the evidence is.
I even believe in Small/Value having a higher expected return than a cap-weighted index. I also believe however, that the level of out-performance of small/value in all those backtest-tools is not likely to be repeated. The difference in CAGR between TSM and SCV is almost 3% a year from 1972-2015! I think it's fair to warn people not to get too excited by all those nice backtested graphs...
Bernstein himself expects the value premium to be much, much lower going forward.

Yes. Getting excited about backtesting at the expense of tactical asset allocation is something that Mr. Browne, Craig R and many others here have truly cured me from - at least most days!

I've posted the link below in another context but I think it might bear re-posting here for a couple of reasons. The author is the principal at one of the most savvy advisory firms through whom one can access DFA funds on a fixed fee rather than % of assets basis. There's a lot of great info on their web site, including the best explanation I've seen on the decisive differences between the DFA funds (which are not index funds) and their counterparts in the retail investing world. The points of interest in this particular article, it seems to me, are looking at how truly diversified equity allocations have performed during market crises, and (further on in the piece) a robust defense of gold that ought to be required reading for a lot of folks over on Bogleheads .

http://www.evansonasset.com/?Page=18

[Dieter]

Tyler actually replaced Small Cap Value with Small Cap Blend in the version of the GB he said he himself has implemented. Aside from the lengthy thread on GB here there are two other equally epic ones on the Mr. Money Mustache and Early Retirement Extreme forums and I don't recall on which of them he mentioned this, but as I recall it was a minor tweak reflecting the fact that TSM already has some value stocks.

Regarding Desert's portfolio, you're referring to "the" desert portfolio featured on Tyler's site, while I'm referring to the one Desert said (earlier in this very thread) he's using now, which is 10% each TSM, SCV, EM and GLD and 60% IT Treasuries. This allocation outperforms the GB in terms of risk:return over any time period I've run the numbers. Unlike GB which is essentially a PP slightly tilted towards stocks this "new and improved" Desert (Desert Oasis? clearly it needs a name to compete with GB ) is basically Larry Swedroe's No Fat Tails allocation spiked with a bit of gold.

Ah, thanks for the clarification. And the listed version of No Fat Tails (https://portfoliocharts.com/portfolio/s ... fat-tails/) more differences as well -- Hard to keep track

But overall, seems to show that less gold and more bond has increased CAGR and reduced Max Drawdown (future mileage may vary, of course.) And aligns with the link Kevin K posted above (http://www.evansonasset.com/?Page=18) :

"Ibbotson found that the risk/return ratio in conservative, moderate, and aggressive equity/fixed portfolios was improved by [Gold] allocations of 7.1%, 12.5%, and 15.7% respectively. "

For GB (SCV version), moving half Gold to LTT and increasing duration slightly gives CAGR of 10.46% ; Max DD of 4.43%
(Vanilla PP for the period: CAGR of 8.4%; Mad DD of 6.46%)

Switching stocks to include EM as per the Desert Oasis split, although 10% more (14% LCB, 13% SCV, 13% EM) gives CAGR of 11.07%; Max DD of 6.83%. But no guarantee that interest rates keep going down....

Given the recent discussion re: Mid-Cap, switching Vanilla PP from LCB to Mid-Cap Blend (same period): CAGR of 9.15; Max DD of 4.61%.

https://www.portfoliovisualizer.com/bac ... 0&Gold3=25

[Edit: Added comment from link posted by Kevin K]

[MachineGhost]

It's post-70's-reset performance has been dismal.

Good to see you back! I disagree, cuz:



It easily surpassed gold's gains from 1975 to 1980.

[Kevin K.]

Good to see you back! I disagree, cuz:

Nope, Desert is correct. When you do the actual comparison he recommended it's very clear that the contribution of gold to the performance vs 60:40 when we came off the gold standard was a once-in-a-lifetime thing.

Here's 1972-2015:

https://www.portfoliovisualizer.com/bac ... 5&Gold2=25

And here's 1975-2015:

https://www.portfoliovisualizer.com/bac ... 5&Gold2=25

[stuper1]

Try comparing say 4-year trailing average real returns for the two portfolios with and without gold. I'll bet you'll find that the one with gold is consistently as good or better.

[Tyler]

An interesting exercise is to go into PortViz and compare a 60/40 stock/bond portfolio with the HBPP, first starting from 1972, then from 1975 (when gold become legal to own). The PP past performance is very influenced by a few years of spectacular gold performance that I never expect to see again.

Sure, but the return of the PP during that timeframe was still within the historical norms. And interestingly, when gold cratered in the 80's the PP did even better than when it skyrocketed in the 70's. That's the cool thing about the PP -- the start date generally has no effect on the overall returns.

[curlew]

Regarding the GB: if something looks too good to be true...
SCV backtests so well because of the very high returns over the last 40 years and because of the low correlation. The holy grail basically.
Both are unlikely to repeat themselves.

Yes, and the PP is also not likely to repeat the same success it has had in the past. Or so I've heard many times.

[koekebakker]

Regarding the GB: if something looks too good to be true...
SCV backtests so well because of the very high returns over the last 40 years and because of the low correlation. The holy grail basically.
Both are unlikely to repeat themselves.

Yes, and the PP is also not likely to repeat the same success it has had in the past. Or so I've heard many times.

To me that sounds a bit like: because some investment did well in the past, it will continue to do so in the future.
I'd like to hear a reason why. In the case of the PP I think there is. I believe it to continue to do well under almost all circumstances.
In the case of SCV I don't see it. Why will it continue to outperform TSM 2-3% a year? And even it does, is it really possible to catch this kind of premium using funds like Vanguard SCV?
I highly doubt it. If there still is a SCV premium it will likely be much lower.
So looking at the backtested results of adding SCV to the PP is looking at a fantasy. Unlikely to repeat itself. It likely won't do any harm, but don't expect the same magic to happen going forward.

[curlew]

Regarding the GB: if something looks too good to be true...
SCV backtests so well because of the very high returns over the last 40 years and because of the low correlation. The holy grail basically.
Both are unlikely to repeat themselves.

Yes, and the PP is also not likely to repeat the same success it has had in the past. Or so I've heard many times.

To me that sounds a bit like: because some investment did well in the past, it will continue to do so in the future.

I'm just applying the same agnostic approach to the GB as we do to the PP. My SCV is up 13% YTD but I have no idea how well it is going to do in the future. Any day of the week I can read articles about how gold, bonds, or stocks aren't going to do well in the future but I've quit paying attention to any of it.

[Dieter]

Regarding the GB: if something looks too good to be true...
SCV backtests so well because of the very high returns over the last 40 years and because of the low correlation. The holy grail basically.
Both are unlikely to repeat themselves.

Yes, and the PP is also not likely to repeat the same success it has had in the past. Or so I've heard many times.

To me that sounds a bit like: because some investment did well in the past, it will continue to do so in the future.
I'd like to hear a reason why. In the case of the PP I think there is. I believe it to continue to do well under almost all circumstances.
In the case of SCV I don't see it. Why will it continue to outperform TSM 2-3% a year? And even it does, is it really possible to catch this kind of premium using funds like Vanguard SCV?
I highly doubt it. If there still is a SCV premium it will likely be much lower.
So looking at the backtested results of adding SCV to the PP is looking at a fantasy. Unlikely to repeat itself. It likely won't do any harm, but don't expect the same magic to happen going forward.

The way I look at it, the big difference in GB is the 40% to stock; 20% each the rest -- tilt to prosperity.

Secondarily is where to put that extra 20% to prosperity. Adding to LCB or Total Stock would work. Or going like Desert Oasis and 1/3 in Totl US, SCV, EM.

Given that TSM / LCB are primarily Large/Huge Growth, I like the added diversification of Small, value, REIT, and/or Intl (Small or EM.)

Couldn't find any of the more in depth analysis, but an overview from a quick google search on why Small-Cap Value might do well: http://www.moneychimp.com/articles/inde ... why_sv.htm

[Mr Vacuum]

The way I look at it, the big difference in GB is the 40% to stock; 20% each the rest -- tilt to prosperity.

Secondarily is where to put that extra 20% to prosperity. Adding to LCB or Total Stock would work. Or going like Desert Oasis and 1/3 in Totl US, SCV, EM.

Given that TSM / LCB are primarily Large/Huge Growth, I like the added diversification of Small, value, REIT, and/or Intl (Small or EM.)

Well said. Browne's principle was to hold the entire market at all times in order to reap the returns of whichever cap/factor/segment is doing well at any given time. Looking past the prosperity tilt, the GB is intriguing because of how explicit it is about doing that, which you don't see that much in a typical 4x25 PP. The morningstar xray with Tyler's 50/50 VTI/VB is a sight to behold with a nearly perfect split across the style box and only two funds.