- 01/17/2015 at 2:03 pm #16016
Use this thread to address more sophisticated questions and suggestions on portfolio composition, optimization and technicalities – so we keep them in one space for other advanced readers.01/17/2015 at 2:04 pm #16017
We received an interesting question by email, we post here to the benefit of the broader audience.
“In several of your blog posts, you advise shorting TMV as a hedge and leaving it in the account. This means we would never realize any gain from it, although we need to pay the cost of borrowing the money monthly. Do you ever sell any of the hedge if the price goes down, to cover the cost of borrowing, and then buy it again when it goes lower?”
Contrary to a long position a short position will lose value over time. So, if you open a TMV short position of 10’000$ and Treasuries go up, then TMV will lose value. After some time your short position will for example only have a value of 9’000$. Now because you want to keep a 10’000$ hedge, you will have to sell more TMV for 1000$.
In fact you are doing exactly the same as someone who has the +3x leveraged TMF to hedge. His TMF position would have gone up from 10’000$ to 11’000$. To keep it at 10’000$ also he would have to sell ETFs for 10’000$.
So the effect of rebalancing is exactly the same for long and short ETFs.01/17/2015 at 2:05 pm #16018
In one of Frank’s comments a while ago I believe he mentioned a minimum single strategy size of $50,000, though on a followup question discussing the BRS he said you could use one etf versus two to keep costs down…..we were discussing a $20,000 account.
If we wish to add Bug to the other 5 strategies (UIS,MY,GM,GS/L,BRS) is there a recommended overall size? Because the strategy holds more than 2 etf’s, perhaps 4-7, then what threshold/overall strategy size would make monthly rebalancing for Bug reasonable. For example, if on average you hold 5 etfs in Bug you may want to start with $10,000 or $20,0000 or more in each knowing there will be monthly rebalancing and transcation fees.
Good job working through this rollout. Thanks.01/17/2015 at 2:05 pm #16019
Thanks for the kind words, and good question Sam. Answer depends how someone is set up with the brokerage . . . Interactive Broker trading fees, or more like $8/trade model? Ultra Low commission and you can make smaller allocations work. Also the personal preferences of time overhead of managing very small trades. Remember, some of the adjustments will be partial shifts, so a 10% allocation shift may mean moving just $1K of a normal $10K position. Finally, there is the tax reporting.
All this depends on your situation and we do not advise individuals, only general ideas. For me, I would work out allocations that targeting trades of at least $2.5 or more per ETF per trade adjustment. See what makes sense for you and project back to how much total allocation that works out to be.
I hope this helps a bit.01/17/2015 at 2:06 pm #16020
Here are three models that offer low volatility and high Sharpe:
Model A B C BRS 30% 20% GMRS 5% 20% 25% MYS 30% 30% 40% UIS 35% 30% 35% CAGR 25% 29% 33% Volatility 9% 11% 13% Sharpe 2.3 2.3 2.301/17/2015 at 2:06 pm #16021
Patrick, thanks, this is great! Let me add the visual representation of these, and we shall call them in future:
Btw, as you might notice what you have done is a portfolio of strategies, or a “Strategy of Strategies”, e.g. treating each underlying strategies’ equity curve as an asset (class) for a next layer of strategies. Currently this is limited to applying fixed weightings, but imagine to create a rotational “Strategy of Strategies” where you invest in the Top 1-3 best performing strategies, or a volatility weighted strategy portfolio where each strategy is assigned or limited to the same “bucket” of volatility, or even a mean-variance or minimum-variance optimization of strategies?
This is what we mean with the “meta-strategy” approach we’re currently building up. Concept is up and running, first tests look very promising, but we’ll take some more time to fine-tune both the concept and the publication to present you with a well-rounded paper. Anyhow, just drop us a line to share your ideas, needs and ideas around this.
[Note these portfolios are not updated to full 2014 data]01/17/2015 at 2:07 pm #16022
Hi Patrick, indeed for our analysis we also use the Sortino Ratio beside the Sharpe Ratio. It depends a bit on the strategy you are evaluating, especially on what risk/return profile you are focusing. For strategies with higher volatility the Sortino Ratio gives a better view, while we prefer Sharpe for lower volatility strategies. Sharpe is the broader known, this is why we use it for publication, but are open for feedback.
The volatiltiy shown here is the annualized monthly volatility, e.g. in Excel StDev(A1:A100) * Sqrt(12). This prototype needs to be light-weight, therefore we did not put daily equity data in. But the result is about the same. This is also the reason we currently do not include MaxDrawDown, as with monthly data it does not reflect the real “fear factor”. The next version will be on daily data level, and with more options to be selected.01/17/2015 at 2:07 pm #16023
When will the “meta-strategy” mentioned earlier in this thread be published?01/17/2015 at 2:07 pm #16024
Michael, thanks for your interest. The ‘meta-strategy’ is part of a major overhaul on the site we’re currently planning, we’re targeting Q1 2015. Will post some examples before to get feedback and develop it to your needs.01/17/2015 at 2:07 pm #16025
Great work by you all. Wouldn’t you want to add correlation as a criteria (both to say S&P500 and each other strategy) then mix the least correlated strategies to further reduce risk?
I’m looking forward to your continuing efforts.01/17/2015 at 2:07 pm #16026
To minimize trading costs consider Folio Investing brokerage. For 1 fixed annual fee you get more window trades than you could ever use with these strategies. Also, you could set up a folio for each strategy within the account for easy strategy management. I have used these guys for years as well as Schwab and IB. Some drawbacks though – no margin and no options.01/17/2015 at 2:08 pm #16027
Excellent point; that is one criteria we look at and I intend to do more in that direction.
Of course, the challenge is developing the strategies that are logical, stable going forward, and non-correlated to each other and the S&P; other than that, it is quite easy. :)01/17/2015 at 2:08 pm #16028
Thanks for the heads up on Folio Investing brokerage. They look like that could work for some subscribers; I will need to give them a call.01/17/2015 at 2:09 pm #16029
Excellent work, just a couple of points
1. portfolios with correlation with each other would be great as there is overlap particularly with regards to the bonds and to lesser degree SPY vs MDY.
2. Curios as to why the US REIT like IYR not included in the sector rotation ETF’s. I note we do have RWX.
thanks01/17/2015 at 2:09 pm #16030
Sunil, thanks for the feedback, appreciated. Agree on the correlation functionality, give me some days to include a correlation matrix between strategies and the “custom portfolio”, this was suggested before.
The emphasis in the “Global Sector Rotation Strategy” lies in “Global”, as the Universe of Sector ETF is quite huge we limited to global indices where suitable. This is why we choose RWX over IYR. We have one strategy on the burner which employs both US and Global REIT, this might be of interest to you.
You must be logged in to reply to this topic.