How to short leveraged ETNs: a backtest on shorting JDST/JNUG
In my previous blog post, I discussed the terrible performance of leveraged ETNs and tried to shed some light on why these products decay over time. The key take aways were:
Even if the underlying moves in favor of the ETN, the ETN might still lose value.
The higher the volatility of the underlying, the bigger the loss of the ETN.
Most leveraged ETNs lose > 90% of their value over the long term.
Just as a brief reminder, the next chart shows the performance of leveraged ETNs on the Junior Goldminers index. While the index itself (grey line) is basically flat over the whole period, both the 3x long and 3x short ETN lose > 90% of their value.
As this was already discussed in more detail in my previous article, this article puts its main focus on how to actually trade leveraged ETNs. The basic trading idea is this: Both long and inverse leveraged ETNs lose value over time, so shorting both of them appears to be a no brainer. There are, however, some issues that reach the surface once we dig deeper into the topic: Directional risk, shorting fees and realized volatility.
Delta Risk (Directional risk)
Delta is usually used to describe the sensitivity of an option to its underlying, but in this context it refers to something different: Shorting both long and inverse ETNs seems attractive, because initially it might appear like there is no exposure to the movement of the underlying. That is only true, however, for the very day the initial trade is conducted. Just like an option trader constantly has to delta hedge his options in order to remain delta neutral, the short ETN strategy would require continuous rebalancing. The reason for this is that the very moment the underlying moves into either direction, one short position becomes bigger than the other short position. To demonstrate this, let me show you how a trade in 3x leveraged ETNs can turn out:
Let’s assume that on day 1, we sell 50k worth of the 3x leveraged long ETN and 50k of the 3x leveraged inverse ETN. If by the end of day 1, the underlying has spiked 10%, the short position of the leveraged long ETN is now worth -65k (keep in mind, the ETN is 3x leveraged) and the position in the leveraged inverse ETN is worth -35k. As a result of that, the trade is no longer delta neutral, since one position is bigger than the other.
It is true that over the long term, both ETNs are expected to approach 0, but on the way there, the investor can be exposed to massive directional risk. Obviously, this risk can be reduced by constantly rebalancing the portfolio, but while constantly rebalancing the portfolio reduces the directional exposure, it also reduces the potential profits that can be made by capitalizing on the negative compounding effect. As a result, traders have to find the right balance between rebalancing to minimize the directional exposure and keeping the positions untouched to capitalize on the negative compounding effect. A trader with a low risk tolerance could theoretically rebalance the portfolio every 2 days, whereas a higher risk tolerance would mean that one rebalancing trade per month (or even year) could be sufficient.
With that in mind, it is about time to preselect leveraged ETNs that ought to be shorted. The previous article has already mentioned that a high level of volatility will make the time decay of leveraged ETNs accelerate and the impact of the level of volatility can be easily witnessed in the trade. The chart below demonstrates the performance of shorting, and constantly rebalancing, JDST and JNUG, 3x leveraged ETNs that aim to track the junior gold miners index (The assumed shorting fees for the trade will be addressed later in the article). In the chart, the performance of the strategy is displayed along with the rolling 30 day annualized volatility of the underlying.
As it should be seen quite clearly, a high level of volatility leads to a massively positive performance, while a decrease in realized volatility, as seen in 2017, negatively affects its performance. The chart below shows again the rolling volatility, as well as the drawdown curve of the very same strategy. It should demonstrate even more clearly, how a collapse in volatility leads towards a negative performance in the strategy and vice versa.
But while volatility certainly decreased in 2017, it still seems relatively high, with an annualized level of mostly >20%. In order to find out why the strategy was still trending downwards, it is important to take a look at the shorting fees of leveraged ETNs.
Once it has been established which ETNs to choose and which to avoid, it is finally time to set up the trade. Opening up a brokerage account and seeing the shorting fees for leveraged ETNs would have left many traders with a headache in the past, as they used to be as high as 16%. An increase in demand over the past years has driven down shorting fees however, and nowadays shorting fees for liquid leveraged ETNs range between 5 and 10% p.a. How big the impact of the shorting fees can be on the overall performance of the strategy is demonstrated by the chart below. It shows the performance of shorting, and constantly rebalancing, JDST and JNUG, 3x leveraged ETNs with 0, 5 and 10% shorting fees. Furthermore, the underwater curve demonstrates quite impressively that particularly in times of low volatility, shorting fees can make the difference between positive and negative performance.
Determining the right balance between shorting fees and volatility, however, is not always easy. In a way, you can think of it in a similar way of thinking of options: Shorting a pair of leveraged ETNS provides positive returns in a high volatility environment (as does buying a put and call option on the same underlying). In order to purchase the pair of options, one would have to pay a premium, whereas to short the ETNs, one has to pay shorting fees. But while the price for the options is indeed influenced by the level of volatility (as it determines the likelihood of making money from a straddle), the shorting fees for the ETNs are only determined by liquidity. This means that very often, highly volatile ETNs are also the cheapest ones to short, as their high level of volatility makes them attractive for both speculators and pure short sellers.
Whether or not to short a pair of leveraged ETNs should not be decided by one of the factors mentioned above, but by taking all of them into consideration. While the delta risk is the same for all ETNs, and can therefore be ignored in the ETN selection process, it is important that the volatility of the underlying is high and the shorting fees are low. Taking all these factors into consideration, shorting leveraged ETNs can be a very lucrative trade that can provide positive returns to investors, as long as all potential impact factors and risks are considered.
Appendix: What could possibly go wrong? The potential tail risk
As such an event has not yet happened, the tail risk has been excluded from the main article, but as anything can happen in the market it is still worth mentioning: With any short position, the maximum return that can be made on a single day is 100%, as the instrument cannot go below 0. But what happens if we short 3x leveraged ETNs and the underlying spikes more than 33% on a single day? The leveraged inverse ETN is pretty straight forward: Most likely, it will reach 0 (as it is 3x leveraged) and we make a profit of 100%. The leveraged long ETN however, can expose us to a problem: if it shoots up in value by more than 100%, every percent above 100 is no longer compensated by the other short position, as it already approached 0. This is the reason why it is important to not just select the leveraged ETN which has the most volatile underlying, but to check, whether the underlying spikes up or down in extreme market movements. Those underlyings that skyrocket when markets tumble, should definitely be avoided in order to stay out of trouble when markets go belly up. Even though I do not really assume that leveraged ETNs are actually able to deliver 3x the return of the underlying in such an event, avoiding such ETNs will probably still make you sleep better at night.
The most prominent example of ETNs that should be avoided for this trade are leveraged VIX ETNs. As the volatility of the VIX is very high, the time decay of leveraged VIX ETNs is enormous and makes them seem like the perfect instruments for this trade. Historical data shows however that there have already been occasions where the VIX spiked by far more than 100% (the biggest spike probably being the market crash 1987), which can leave the trader in a very unfavorable position. As the maximum leverage of VIX ETNs is 2x and the ETNs aim to track the SPVXSTR index, rather than the VIX itself, it is unfortunately not possible to see how ETNs actually behave in such an extreme event. But even though I do not really assume that these ETNs actually delivering that kind of performance in such an event, avoiding such ETNs will probably still make you sleep better at night.
Also published on Seeking Alpha: https://seekingalpha.com/article/4301463-short-leveraged-etfs-backtest-shorting-jdst-jnug