The Golden Butterfly Portfolio



The Golden Butterfly Portfolio

The Golden Butterfly originated from Tyler’s  Portfolio Charts and is often considered as a derivative of Harry Browne´s Permanent Portfolio which I looked at in a previous post.  Like the permanent Portfolio is comprises of 4 asset types:-

  • Equities
  • Short term government bonds bonds
  • Long term government bonds
  • Gold

THE GOLDEN BUTTERFLY

Whereas the Permanent Portfolio has equal weightings of each asset,  the Golden Butterfly aims for higher growth by allocating 40% to equities split evenly between Total Market and Small Cap Value and equal allocations to short term government bonds, long term government bonds and gold.

As with the Permanent Portfolio the idea is to have four asset classes to deal with all likely market conditions  - equities provide growth in expansionary markets, gold  in inflationary markets, long term bonds during deflationary periods bonds and short term government bonds/cash provide protection during market crashes.  

ASSET CLASS PERFORMANCE

The Golden Butterfly provides more growth potential by allocating 40% to equities rather than 25% and an allocation of 50% of the equities to Small Cap Value (see Paul Merriman for more information about Small Cap Value).  The higher equity allocation results in higher volatility than The Permanent Portfolio but an average growth rate more than 2.5 percentage points higher. 

Many people will be wary of this portfolio.  Gold is not a well liked asset class and many investors will be wary of investing in long term bonds with interest rates at historic lows.  A further problem is that there are only around 50 years of data available on some of the asset classes so we have no knowledge of how the portfolio would have performed during the late 60s and early 70s when there were 2 market crashes close to one another.  However, the out performance of this portfolio and the rational behind the asset allocation makes it a very attractive option.

My Golden Butterfly Portfolio


GOLDEN BUTTERFLY PORTFOLIO
ASSET
%
CASH
20%
Ishares Gilts 0-5yr
TOTAL STOCK MARKET
20%
iShares Core MSCI World ETF
SMALL CAP UK
10%
BlackRock Throgmorton IT
Standard Life UK Smaller Companies 
SMALL CAP GLOBAL
10%
Edinburgh Worldwide IT
Smithson IT
LONG TERM GILTS
20%
Vanguard Long Duration UK Gilt Fund
GOLD
20%
iShares Physical Gold ETF

   
I have chosen 4 Investment Trusts for the 20% small cap segment of the portfolio, 2 UK based and 2 global and have taken a risk in choosing Smithson but I did the same when Fundsmith launched and that gamble certainly paid off.  Opting for 4 trusts to represent the small cap sector makes portfolio balancing more complicated but reduces the risk of being dependent upon one trust and one manager but without doubt more monitoring will be required than for an index fund.

 So this is how my final portfolio looks (performance as of 11/3/2020):-

Manager
Unit Name
% of Portfolio           
1 Year Perf.




iShares
20.0
-1.5
Vanguard Investments UK 
20.0
28.8
iShares
20.0
1.8
iShares
20.0
27.8
Fundsmith
5.0
6.4
Baillie Gifford & Co Ltd
5.0
1.2
Standard Life Investments
5.0
14.3
BlackRock Investment Management
5.0
12.3
Total Portfolio
Total Portfolio
100.0
12.7




Selected Benchmark
FTSE All Share

-10.7

As one would expect the gold and long term gilts helped protect the portfolio from the Feb/March 2020 market declines.   


Five year performance is shown below (without Smithson due to its newness compensated for by an  extra contribution from Edinburgh Worldwide):-

Manager
Unit Name
% of Portfolio
5 Year Perf.




iShares
20.0
48.5
Vanguard Investments UK
20.0
62.9
iShares
iShares UK Gilts 0-5yr ETF GBP
20.0
5.2
iShares
20.0
63.6
Baillie Gifford & Co Ltd
10.0
105.6
Standard Life Investments
5.0
89.6
BlackRock Investment Management
5.0
110.8
Total Portfolio
Total Portfolio
100.0
50.9




Selected Benchmark
FTSE All Share
11.6



The portfolio has higher growth and lower volatility than the FTSE All Share:-

Performance Analysis
Growth Rate %
Volatility
Alpha
Beta
Sharpe
Info Ratio %
Portfolio
8.8
6.5
7.4
0.4
0.8
0.6
Benchmark - FTSE All Share
3.6
10.7
0.0
1.0
0.0
0.0


International Diversification

Monte Carlo Simulation

A 5 year performance history is a good indicator of relative performance but only tells a small part of the story.  A Monte Carlo simulation is more useful as it analyses a portfolio of similar asset classes over a large number of historical and simulated scenarios.  I have used the Timeline simulation tool to compare the Golden Butterfly with a conventional 60/40 and permanent Portfolio. An exact replication wasn´t possible as Timeline do not have long term gilt or intermediate gilt asset classes, their nearest approximation being UK Aggregate Bonds.



ASSETS
60/40
PERMANENT PORTFOLIO
GOLDEN BUTTERFLY
Cash

25%
20%
UK Aggregate bonds
40%
25%
20%
UK Equities
60%
25%

UK Small Cap


10%
Global Equities


20%
Global Small cap


10%
Gold

25%
20%

The simulation is based upon a 4% withdrawal rate, 3.8% inflation and a target 30 year retirement. The simulation gives the Success Rate in not running out of money before the 30 years have ended, the age at which 10% or fewer retirees would run out of money (assuming an age 65 retirement start), the minimum legacy that 9 our of 10 retirees would leave and finally the lifetime income received.

The Golden Butterfly significantly outperforms the other two portfolios.



OUTCOMES
60/40
EQUITY/BOND
PERMANENT PORTFOLIO
GOLDEN BUTTERFLY
Success Rate
88%
83%
98%
Longevity
93 years
92 years
95 years
Legacy
£0
£0
£278.2k
Lifetime Income
£1.13m
£1.12m
£1.2m

  
Another useful means of comparing performance using historical actual data is through Portfolio Charts.  Again the exact asset classes are not available but using the best approximations does provide a means of comparing portfolios. 


  • Max. Drawdown: Maximum decline in value
  • Recovery Time. How long to recover from a downturn
  • Average Return: Average Inflation Adjusted Return
  • Safe Withdrawal Rate: The initial withdrawal rate that ensures not running out of money during a 30 year retirement with annual withdrawals increasing by the rate of inflation
  • Perpetual Withdrawal rate: The initial withdrawal rate that ensures never running out of money with annual withdrawals increasing by the rate of inflation
PORTFOLIO CHARTS COMPARISON

Max. Drawdown
Recovery Time
Average Return
Safe Withdrawal Rate
Perpetual Withdrawal Rate
60/40 Equity/Bond
57%
12yrs
6.1%
4.0%
3.2%
Permanent Portfolio
9%
5yrs
4.7%
5.3%
3.7%
Golden Butterfly
22%
5yr
5.3%
5.7%
4.3%

The Golden Butterfly permits the highest Safe and Perpetual Withdrawal rates at 5.7% and 4.3%, with a much reduced maximum drawdown and faster recovery time than a 60/40 portfolio.  The Permanent Portfolio offers the lowest maximum drawdown but lower withdrawal rates than the Golden Butterfly.

Conclusion

Using both historical data and Monte Carlo Simulation the Golden Butterfly is the top performing portfolio for retirement drawdown.   There are some caveats: some historical data is limited, some investment products have been in existence less than a decade and whilst is is relatively easy to simulate a US portfolio, the UK investor can only approximate to the ideal Golden Butterfly due to the absence of a UK Small Cap value class and of course finally - the past is no accurate predictor of the future.

Post a Comment

8 Comments

Unknown said…
Really interesting article, I was looking for some starting points for a UK variation having found the US original on portfoliocharts.com I look forward to an update on the anniversary.
Max said…
Glad you found the post useful. I´ll be posting an update in a few week's time. I´m currently researching the effect of gold on UK portfolio volatility and SWR - and will be posting next week - if the Golden Butterfly has drawn your attention you will find it interresting.
Gordon said…
I'm looking at possibly implementing a version of this portfolio in my SIPP. I plan to base my withdrawals on the ideas in Michael McClung's book.

For the purposes of implementing McClung's Prime harvesting strategy and EM withdrawal strategy, would you attribute the Gold allocation in this portfolio to the equities or to the bonds side of the equation?
Max said…
Hi Gordon I think that´s a really good question. I hold gold in all of my portfolios (12.5%-20%) - but I decided not to implement Prime Harvesting because variable withdrawal strategies seem to offer such an improvement in income potential that Prime Harvesting benefits are insignificant in comparison. Perhaps, more importantly, is that my portfolios are all similar to the Golden Butterfly and all my analysis is based upon maintaining the balance between the asset classes so Prime Harvesting would not be applicable.

But if I were to implement Prime Harvesting with gold within the portfolio I would wish to keep the % of gold pretty well fixed so if the portfolio were equities, bonds and gold I would initially take gold out of the equation playing with just the equities and bonds and selling gold only if it exceeded significantly my target - in that scenario, you could end up with an equity/gold portfolio at some stage but historically that would have performed quite well.
Gordon said…
Thanks Max,

That's pretty much the conclusion I was leaning towards myself. I agree with your comments about Prime Harvesting, I've been running my own Monte Carlo simulations in Excel, and I've noticed that on quite a high percentage if iterations, Prime Harvesting actually does very little, certainly in comparison to the benefits of the variable withdrawal strategies as you say. However it's difficult to know whether under certain market conditions it might just make the difference between running out of money or not.

It took me a while to realise the true value of having gold in a portfolio, until I read Tyler's excellent article on the Portfolio Charts website - (https://bit.ly/NoiseCancellingPortfolio) and suddenly it all fell into place - and I'm now a convert.

Interesting to note that the Golden Butterfly does very well indeed, when using McClung's Harvesting Ratio metric - reinforcing the claims that it supports higher withdrawal rates than many other lazy portfolios.

I found your blog to be an interesting read, and hope you will be posting more articles in the future.
Max said…
Hi Gordon - I´m very pleased you´re enjoying the blog - it was my research and analysis for today´s post about Variable Drawdown that convinced me not to bother with Prime Harvesting. You seem to have gone down a very similar research route to me and the McClung book is very impressive in its thoroughness. Coming to a firm decision about gold isn´t easy as history is limited (it doesn´t work well on Timeline due I think to them using data back to around 1900 so results are distorted by 70 years of gold price fixing)) but I guess we need to trust that it will continue to be strongly negatively correlated with equities so should be valuable in reducing portfolio volatility.

I think one of the challenges of retirement finance is that whilst it is fascinating to pursue the optimum strategy it may be necessary to compromise in order to have a plan to be easily be implemented by someone else - partner, relative, financial adviser should we end up mentally incompetent - or just bored with finance!
Unknown said…
Any update on this please? Looks very very interesting to me .
Would be extremely gratefulan for an update .
Max said…
I’m waiting for the next market crash to see how it performs. Hopefully not too soon!