Quantopian's community platform is shutting down. Please read this post for more information and download your code.
Back to Community
Shorting 3x ETFs with regular rebalancing

Is this a viable algo for trading? It looks like very high Sharp and low drawdonws. Please comment

I used the algo posted here as a source and just slightly modified it :https://www.quantopian.com/posts/is-this-a-good-rebalance-algorithm-or-does-it-need-work

5 responses

Shorting leveraged ETFs is not without risks; you are essentially selling gamma, so you are short volatility, and must rebalance to avoid getting overexposed. At 30 day rebalancing, you'll be pretty vulnerable to sustained moves in natural gas. The LETFs have short fees that range from 4% to 16%, and can quickly become difficult to borrow. You are also vulnerable to the ETN sponsor halting redemptions or creations, which will cause an persistent distortion in one of the legs, making the arbitrage impossible, until they resume. This happened to TVIX a while ago.

There's a ton of nonsense about these strategies on Seeking Alpha. Here's some counterbalance:

http://seekingalpha.com/article/3140956-investing-in-leveraged-etfs-theory-and-practice

Incidentally, this is related to the strategy I won the contest with, which did quite poorly on August 24th. Don't let me dissuade you, but I am pretty skeptical of this strategy these days.

Thank you for the feedback. Reducing the re-balancing to 10 days decreases overall return from 200% to 50%. It's mostly due to the fact that the longer the period between the re-balancing the more time decay inherent to leveraged ETFs can be captured.

Maybe re-balancing based on threshold deviation from optimal portfolio will be optimal from risk the management point.

Overall, I think that this strategy can be useful diversification tool in a portfolio because of its low Beta and positive expected return.

The decay is only when the volatility outweighs the sustained returns, so the positive expected return is contingent on the future behavior of the market. Caveat emptor!

Caveat emptor indeed!

EDIT: actually backtest looks like the price split problem so not indicative of algo in anyway
https://www.quantopian.com/posts/ugaz-is-not-udjusted-for-split

If you rebalance at any frequency other than daily like the underlying portfolio of the ETF then you're going to have periods with big drawdowns. Its purely mechanical based on how much each leg grows or shrinks due to the rebalancing and the subsequent movement of the index. Some 3X ETFs are inefficient in their internal rebalancing so you can take advantage of that by rebalancing the ETFs themselves daily. However as Simon points out, the short fees far outweigh the slow and steady upward drift you get from the inefficiency. This is one of many apparent arbitrages that work in the frictionless world of theory but can't be actually implemented. Also don't forget that this is something you'll have to just do in your personal portfolio, as Quantopian has been consistently adamant on their feelings about leveraged ETFs .