I was doing a little research over the holidays when I encountered an interesting post on the Subreddit WSB. Unfortunately, WSB is now private, so it would be pointless to share the link, however, it did lead me to dig deeper into understanding leveraged ETFs; specifically how they’re structured and what types of drag are involved. I found this post that goes into detail on the “volatility drag” that leveraged ETFs face, but an optimal leverage ratio greater than 1 can be obtained. It should be noted that this optimal leverage ratio seems to only apply when the ETF is going long on its underlying benchmark; leveraged inverse ETFs seem to face a decay over time, which I believe has something to do with its volatility drag being exacerbated by its path dependency (path dependency is explained in more detail in this blog).

With this in mind, there were three ETFs I looked into that were the best in my opinion to take advantage of a short position:

- TZA – Direxion Daily Small Cap Bear 3X ETF. Note that since inception, this ETF has lost 99.89% of its value.
- FAZ – Direxion Daily Financial Bear 3X ETF. Similar to TZA, it’s lost 99.95% of its value since inception.
- UVXY – ProShares Ultra VIX Short-Term Futures. This ETF is actually going long on the VIX, in which, the mechanism for its decay is actually due to contango in the VIX futures term structure. However, because of its leverage, the volatility drag exacerbates the decay for this instrument. UVXY has lost 99.99% of its value since inception.

Below, I’ve put together a table of return summary statistics for the mentioned ETFs:

Statistics | TZA | FAZ | UVXY |

Total Occurrences | 1315 | 1315 | 1315 |

Mean | -0.23% | -0.25% | -0.63% |

Std | 3.34% | 3.14% | 8.05% |

Min | -17.76% | -17.67% | -26.66% |

25th Percentile | -2.23% | -1.89% | -5.05% |

50th Percentile | -0.36% | -0.39% | -1.15% |

75th Percentile | 1.59% | 1.34% | 3.04% |

Max | 14.28% | 15.38% | 43.75% |

Since all three ETFs are bearish, I’ve calculated their daily return correlations:

TZA | FAZ | UVXY | |

TZA | 100% | 88% | 76% |

FAZ | 88% | 100% | 78% |

UVXY | 76% | 78% | 100% |

We see that TZA and FAZ are highly correlated, so I decided to narrow down my ETF choice to just TZA. With this decision, I chose IWM to use for a momentum indicator since TZA is tracking 3 times the inverse return of the Russel 2000 index. I then optimised the SMA cross against Sharpe ratios:

Which gave a 55-110 SMA cross for IWM to be an optimal signal. I then applied the ratio and backtested with TZA from 10/04/2011 to 12/23/2016:

We see that the results are pretty decent. For UVXY, the same VIX/F1 signal can be applied for entry. Running a backtest, we see that the signal works very well for UVXY:

Now, what does this mean for trading strategies? For one, when dealing with high leverage ETFs, I’d rather not have too much exposure to them, but the results from above are too good to ignore. Thus, to take advantage of this, I propose buying medium term puts on TZA and UVXY following their respective signals and closing positions out when they’re profitable (but since I don’t have access to options data, I can only guess on when to close). This way, your max loss would just be the cost of your puts, which is completely decided by you.