I think I need to retract my previous message. I'm not sure that it's all wrong, but I'm no longer confident in what I wrote, because, checking DWTI (negative 3X OIL ETF), I see the following movement from 2015-12-04 to yesterday's close:
DWTI: $148.50 --> $57.71 (down 61.1%)
The speed of volatility drag, according to Zhang's PhD thesis, and according and given the assumptions he described, should be proportional to (L**2)-L (where L is the leverage factor, which would be 3 for a 3X ETF like UWTI and -3 for a negative 3X ETF like DWTI). This expression is 6 for a 3X ETF and 12 for a -3X ETF. So, the decay rate of a -3X ETF should be twice as fast as that of a 3X ETF, so the decay of a -3X ETF should be about the square of a 3X ETF.
However, if I attribute only a 10.9% drop in the value of UWTI to that decay, the decay from rebalancing of DWTI should be 20.6% (that is, 100% - (100% - 10.9%)**2), but DWTI dropped much more than that.
I suspect that part of the apparent disagreement between theory and reality here is that perhaps due to the difference between the actual movement of oil futures prices and the random walk assumed in Zhang's thesis. Intuitively, perhaps the daily prices have been more mean reverting.
DWTI: $148.50 --> $57.71 (down 61.1%)
The speed of volatility drag, according to Zhang's PhD thesis, and according and given the assumptions he described, should be proportional to (L**2)-L (where L is the leverage factor, which would be 3 for a 3X ETF like UWTI and -3 for a negative 3X ETF like DWTI). This expression is 6 for a 3X ETF and 12 for a -3X ETF. So, the decay rate of a -3X ETF should be twice as fast as that of a 3X ETF, so the decay of a -3X ETF should be about the square of a 3X ETF.
However, if I attribute only a 10.9% drop in the value of UWTI to that decay, the decay from rebalancing of DWTI should be 20.6% (that is, 100% - (100% - 10.9%)**2), but DWTI dropped much more than that.
I suspect that part of the apparent disagreement between theory and reality here is that perhaps due to the difference between the actual movement of oil futures prices and the random walk assumed in Zhang's thesis. Intuitively, perhaps the daily prices have been more mean reverting.
