Here's a said story: https://www.reuters.com/article/us-...suicide-after-vw-losses-idUSTRE5055O820090107 huge short squeeze in volkswagen during the porsche takeover saga. This billionaire was short the normal and long the preferred shares... quite a common trade at the time, since both had the same dividend payout but prefs were quite a bit lower. Prefs didn't have voting rights though.. Porsche was buying the normal shares... spread widened prefs at 110 and normal at 200+. Massive short squeeze ensued, because of forced buy-backs due to short-lending issues, causing the normal shares to hit 1000 in 2 or 3 days. This guy killed himself after his investment vehicle was forced to close the position (forced buy back) and his company, I think it was Heidelberg Cement, was getting into trouble due to bank loans etc.... I think he was still a billionaire, but maybe couldn't face it that he lost 80% and that he was about to lose his life long business. Moral of the story... if you think the impossible would never happen, guess again.
Now, hold there sailor (or is it pilot?) You can mathematically prove that leveraged ETFs are inherently short convexity, I have even derived it for someone. The path dependency, while there, is secondary - it's equivalent to pointing out that because of gamma weighting, realized volatility sensitivity for an option is path dependent.
The contango and 2x part of 2X vix etf both have serious decay effect. Of course when vix suddenly raises a lot, backwardation and 2x part can let it raise much faster than just 2x.
True on VXX. I haven't thought through it but if you were shorting VXX to capture the contango decay how is that any different from selling the VIX futures? You'd expect the same constant profit with occasional blowups that wipe you out, so even though it's a seemingly foolproof strategy over the long term, the amount of capital you have to reserve for when the VIX enevitably jumps gives you an overall very poor return.
Looking for your derivation... aren't we saying the same thing though. Essentially a leveraged ETF is a bet on volatility since it's "decay" is proportional to the volatility of the underlying? Perhaps path dependency isn't the best term, what I'm really saying is that it all boils down to the percentage daily return feature of the leveraged funds. If you do a backtest where you short a 2X fund and go long twice the dollar value of a 1X fund that tracks the same underlying, and rebalance every day to ensure that you still have exactly twice the dollar value in the 1X fund as the 2X fund to eliminate the impacts of daily percentage return, you see a very gradual profit as you would expect from the slightly greater inefficiencies of the 2X fund. If you do it with two funds from the same company that internally balance or use derivative instruments they buy from GS, it stays remarkably flat. Of course you can't execute it in real life because of transaction and borrow costs, but it does demonstrate that the majority of what so many call "decay" in leveraged funds is really just a mathematical result of the percentage daily return design of those funds. In any event, if we are talking about different things then decay from internal fund inefficiency would be a third order effect! Oh and just to check we are talking general daily percentage return leveraged ETFs, not VXX? Volatility of volatility always requires a little extra thought for me.
It's similar. VXX is more constant I think, and you don't have to roll the futures. VXX is basically a constant 30 DTE VIX future... so it rolls the front future over to the next every day to stick to the 30 day maturity. Theoretically, if you just short the front VIX... I would think that the decay picks up speed near the expiry... but it also leaves you with more volatility, since at say 5 days until maturity the vix future would move more than VXX since it's 5 days vs 30 days. I haven't thought out whether it's cheaper or not to do VIX instead...