Leveraged ETF risk parity portfolio

[Ian_944]

Introduction

While looking for trading strategy ideas I came across an excellent site which provided an interesting example of combining leveraged ETFs and a risk parity portfolio management approach for some impressive performance. I repeated and updated the results, and to whet your appetite, here is the performance graph for the period 4-Dec-2012 to 15-Dec-2017.

 

 

There are various good resources explaining risk parity portfolios (the link above and here), but in summary a risk parity portfolio is constructed by selecting two or more inversely or uncorrelated instruments, then investing in these instruments in proportion to the amount of risk (e.g. standard deviation of returns) contributed by that instrument. The portfolio is rebalanced from time to time in order to keep the proportions correct. In a simple example, a portfolio is constructed and initially holds 80% bonds and 20% equities, the low equity weighting is driven by the higher volatility of equities compared to bonds. Going forward, if volatility ('risk') of equities drop over time relative to bonds, the fund will get rebalanced to increase exposure to equities. If volatility of equities increases (e.g. a rapid rise or drop) relative to bonds, then the fund will rebalance out of equities and into bonds. 

 

The issue with a risk parity portfolio is that exposures to high risk/high return instruments like equities will be limited and as a result returns, while more stable, will be lower than required by investors. Hedge funds address this issue by applying leverage to the risk parity portfolio, which obviously is not so easy for a retail investor, but spread bets open up some interesting options in this regard.

 

Additionally, there is a well-documented issue of using leveraged ETFs as long term investments as their daily rebalancing can significantly reduce their performance, particularly in high volatility environments. A risk parity portfolio approach should mitigate this issue by reducing exposure to the instrument as its volatility increases.

 

The Portfolio

The graph at the top of the page shows the performance $10k invested in a portfolio constructed from the ProShares UltraPro S&P500 3x leveraged ETF and the Direxion Daily 20+ Yr Treasury Bull 3x leveraged ETF.

 

The portfolio was rebalanced regularly using a risk parity strategy. The sub-graph below the performance graph shows the proportion of the portfolio invested in the Proshares ETF, with the balance obviously invested in the Direxion ETF.

 

I think the phenomenal performance in the last year is more a reflection of the very unusual market conditions that have existed over the last 12 months, with record low volatility and a steady grind higher in prices (goldilocks conditions for leveraged ETFs), than it is a reflection of the risk parity strategy.

 

The Experiment

While the historical performance speaks for itself, I want to run the portfolio for a time in a demo account for the following reasons:

• Shake down the Python code that monitors and rebalances the portfolio
• Measure the portfolio performance and cost when owned as a spread bet versus in a stock trading account, and also whether a daily funded spread bet or future spread bet is more efficient

I have set up a portfolio with an initial £20k notional value a spread bet demo portfolio, and I will publish its performance, costs and the rebalancing requirements on a weekly basis for the next few months.

 

I feel like the above is a disjointed and probably leaves a lot unsaid, but given it's the Xmas period and I've got family and dogs running around, it will have to do for the moment.

 

Feel free to ask questions!

 

Ian

 

 

[Ian_944]

Here is a summary of the first week of performance for the demo portfolios (up to Fri, 15 December).

 

[Ian_944]

As of 23 December.

 

 

[Ian_944]

There have been some fun gyrations in the US bond market over the past two weeks due to the passing of the tax bill, and it is interesting to note how this additional volatility, along with a quiet stock market, has pushed the optimal portfolio to want to be heavily weighted towards stocks (via UPRO) as we head into 2018.

 

 

And finally, here is a graph showing the full year 2017 performance of the strategy.

 

All the best to everyone for 2018!

Ian

 

 

[Ian_944]

I haven't updated charts in a while as the market was more than boring in early Jan, however the recently relative increase in volatility in SPY makes it worth updating on the portfolio.

 

Unsurprisingly as SPY makes new highs, the portfolio makes new highs. But what has changed notably is that the increase in volatility has for the first time caused the algo to want to reduce exposure to SPY. We are currently overweight SPY vs an ideal balance, not quite at the point where a rebalance is called for, but getting there.

 

Another interesting point to note is how expensive the spread on the March contract looks to have been compared with the accumulated DFB costs which are running behind what I would have expected given the hold period to date.

 

 

 

 

 

[JamesIG]

Interesting thread! 

[Ian_944]

Well, what a week on the US markets! Simply a rocket ship, but the trend of increased volatility continued this week and has caused the strategy to trigger a rebalance, it actually fired on Tuesday, but as I only update over the weekend I will rebalance Monday.

 

The table below shows that the optimal portfolio mix needs to be 55% S&P500 and 45%Treasury, which means I will need to reduce the position (bet) size for S&P500 to 0.71 and increase Treasury to 4.96, which will move roughly £2.2k out of S&P500 and into the other fund.

 

As an aside, if I was running this portfolio in real life I would be very tempted to scale the whole position down to take some profit given our return on capital is sitting at almost 50% in a little under two months.

 

 

 

 

 

[Ian_944]

Unsurprisingly the market action from the past two weeks put a significant dent into this portfolio. While a long term holder would still be comfortably ahead of the game, the demo trading account I have tracked since December has swung from a +10% return to being down by -10% in terms of notional value.

 

The rebalance from a few weeks back helped save a couple percent, but was a drop in the ocean given how rapid volatility ramped. Last week another rebalance was triggered and the portfolio is currently running at the lowest equity exposure in its history with only a 28% exposure to equities.

 

Graphs below if you enjoy visuals of carnage

 

I have been thinking about how best to use leveraged ETFs in light of the performance of this portfolio, as well as the silliness that is the XIV situation. XIV was always doomed to go to zero (and how that was lost for some folk is lost on me), but that issue aside I have been thinking about the warning on all the prospectus's which highlight they are for use as short term trading vehicles only.

 

I'd argue it's not completely accurate to say they are short term only, as both this "risk parity" portfolio and even a terminal product like XIV have exhibited periods where owning them for periods far in excess of their design has yielded strong performance in the right conditions.

 

I think a more accurate statement would be to say leveraged ETFs should not be used in a manner that assumes that returns will compound, e.g. avoid a pure buy and hold approach. And actually if you look at the math and path dependency in these products, it feels intuitive although I have not proven that out.

 

So what to do?

 

I think the solution to utilising these products optimally over the long term is to fix the size of the investment. For example, using my risk parity portfolio as an example: initially invest to a notional value of £20k, but at regular intervals sweep profits out of the portfolio to maintain that initial value of £20k. e.g. portfolio goes up to £21k in a month, then sweep £1k out of the portfolio to bring it's notional value back to £20k. 

 

I need to model exactly how this would play out, but intuitively the result would be a reduction in the parabolic outperformance shown in the graph below, but also a dramatic reduction of any draw downs. 

 

To be precise: I think my next generation effort for this portfolio will do the following:

• Run the same core investments and volatility-based rebalance strategy
• Add an additional rebalancing overlay that sweeps profits in excess of the initial investment into a short duration treasury ETF such as SHY. (e.g. sweep every time profits exceed 10%)

 

 

Given that this thread has generated little interest, I do not plan on making further updates.

 

Regards,

Ian

 

 

 

 

[JamesIG]

Hi  - thanks for the update, especially over the recent market volatility. Leveraged ETF's can certainly be a tricky to trade, and it's important to know what you are getting into before execution. We have an https://etfscreener.ig.com which you may be interested in, and if you follow the link you can download fact sheets which show the constituents and weightings which make them up. 

 

I appreciate when you write a post and there isn't much interaction it can be a bit disheartening, however this thread has been opened over 300 times and I know that the majority of people will happily look at a thread, love whats posted, but not comment or interact at all. So this is a thank you for your insightful contributions!  

 

EDIT: further to this it has also been featured which pulls the post into the Community MyIG widget as shown below (bottom right). 

 

[Caseynotes]

Hi , very much agree with  here, your posts make very interesting reading and though the topic is too far outside my scope to feel comfortable commenting on I very much appreciate the work put in and the updates you have produced.

 

The last couple of weeks have been a real upset to the general flow of many assets classes but corrections happen and how a plan deals with them provides insight to possible changes as you have shown and are experimenting with.

 

     

[Ian_944]

Thanks guy - I appreciate the feedback.

 

Ultimately the thread has served its purpose for me in getting a good sense of how these products behave in a portfolio, and getting a practical understanding of the funding costs associated with owning on margin.