3 ms·

I mean leverage and inverse ETF's aren't broken, they are behaving as desired. You just should never, ever hold one overnight as per the SEC guidelines due to

I mean leverage and inverse ETF's aren't broken, they are behaving as desired. You just should never, ever hold one overnight as per the SEC guidelines due to how they reset each night. https://www.sec.gov/investor/pubs/leveragedetfs-alert.htm https://www.sec.gov/investor/pubs/leveragedetfs-alert.htm I can't think of a single valid reason to hold a leveraged ETF overnight. If anyone has one, please let me know:) The awesome KidDynamite blog goes into how these ETF's are re-balanced and how they can "rip your face off" if you don't understand them. http://kiddynamitesworld.com/leveraged-etf-trading-flow-case-study-gold-miners-gdx-nugt-dust/ http://kiddynamitesworld.com/leveraged-etf-trading-flow-case... http://kiddynamitesworld.com/leveraged-etfs-and-high-school-level-mathematical-truths/ http://kiddynamitesworld.com/leveraged-etfs-and-high-school-...

fwdpropaganda 9y ago
Hey, awesome links, thank you. I was just thinking that I should eventually get around to understading leveraged ETFs, XIV and such. And I end the day with an awesome new addition to my list of RSS ;-)
MuffinFlavored 9y ago
why to hold overnight? look at what XIV did the past 18 months and compare it to SPY :)
Donald 9y ago
This wasn't overnight decay (plus this is short vol, so contango works for you.) This was panic selling as XIV neared its terminal drawdown point.
snthpy 9y ago
Nothing too mysterious in the examples in the SEC Alert link. One just has to keep in mind that these ETFs leverage the daily return and then take into account the compounding effects. If you have an index with normally distributed returns with mean m and volatility v, then the long term compound return c is c = m – 0.5 v^2 . An ETF which leverages k times will have returns with mean km and volatility kv. The ETFs long term compound return e will be e = (km) – 0.5 (kv)^2 = k m – k^2 0.5 v^2 = k (m – 0.5 v^2) – (k^2 – k) 0.5 v^2 = k c – k (k – 1) 0.5 v^2. For k > 1, the second term will always be negative so the ETFs return will always be less than the leveraged compound return of the underlying index. Whether the leveraged ETF return will be more or less than the underlying index return depends on whether c > (k – 1) 0.5 v^2 <=> m – 0.5 v^2 > (k – 1) 0.5 v^2 <=> m > k 0.5 v^2 <=> m / (0.5 v^2) > k. In other words, the leverage k has to be kept less than the m / (0.5 v^2) which is the well known Kelly Criterion.

fwdpropaganda 9y ago

Hey, awesome links, thank you. I was just thinking that I should eventually get around to understading leveraged ETFs, XIV and such. And I end the day with an awesome new addition to my list of RSS ;-)

MuffinFlavored 9y ago

why to hold overnight? look at what XIV did the past 18 months and compare it to SPY :)

Donald 9y ago

This wasn't overnight decay (plus this is short vol, so contango works for you.) This was panic selling as XIV neared its terminal drawdown point.

snthpy 9y ago

Nothing too mysterious in the examples in the SEC Alert link. One just has to keep in mind that these ETFs leverage the daily return and then take into account the compounding effects. If you have an index with normally distributed returns with mean m and volatility v, then the long term compound return c is c = m – 0.5 v^2 . An ETF which leverages k times will have returns with mean km and volatility kv. The ETFs long term compound return e will be e = (km) – 0.5 (kv)^2 = k m – k^2 0.5 v^2 = k (m – 0.5 v^2) – (k^2 – k) 0.5 v^2 = k c – k (k – 1) 0.5 v^2. For k > 1, the second term will always be negative so the ETFs return will always be less than the leveraged compound return of the underlying index. Whether the leveraged ETF return will be more or less than the underlying index return depends on whether c > (k – 1) 0.5 v^2 <=> m – 0.5 v^2 > (k – 1) 0.5 v^2 <=> m > k 0.5 v^2 <=> m / (0.5 v^2) > k. In other words, the leverage k has to be kept less than the m / (0.5 v^2) which is the well known Kelly Criterion.