What killed TBT was that 20% in the middle of the year. It never completely recovered after that. This outlines the decay problem with these leveraged ETF. They make for nice swing trading but are dangerous for long term investment.
Showing posts with label TLT. Show all posts
Showing posts with label TLT. Show all posts
Tuesday, January 11, 2011
Another Take on TBT-TLT
Another take on the TBT / TLT relationship. In yesterday's chart I was showing that the relationship between TBT and TLT on a long term basis was a catastrophe waiting to happen. But if we look at relative performance through 2010, it might actually be worse than that! While TLT is basically flat for the year, TBT is down 25%!
What killed TBT was that 20% in the middle of the year. It never completely recovered after that. This outlines the decay problem with these leveraged ETF. They make for nice swing trading but are dangerous for long term investment.
What killed TBT was that 20% in the middle of the year. It never completely recovered after that. This outlines the decay problem with these leveraged ETF. They make for nice swing trading but are dangerous for long term investment.
Monday, January 10, 2011
TLT Weekly Chart
Below is a chart of TLT on a weekly basis. The red circles outline the selling points - on the first one, my long term indicator and Parabolic SAR went negative. On the second one, my short term indicator went negative as well which started the big downtrend of the last couple of months. However, the green line outlines a good support line around $88 that has held since 2008. We could see a bottom soon, but the long terms prospects are not good as rates will have to go up either because the Fed decides it or investors will make the decision for them.
Tracking TLT with TBT - 3 Years
After running the system on 3 years of data, it appears that using TBT to track TLT over the long run is a fool's errands. Volatility and decay make it almost impossible to create a working model. The best result I can obtain using some overly complicated formulas is displayed below:
And here are the stats:
As time goes by, decay kills the model and will do the same to your investment! Better to stick to TLT. TBT makes a nicer vehicle for short term trades though!
As time goes by, decay kills the model and will do the same to your investment! Better to stick to TLT. TBT makes a nicer vehicle for short term trades though!
TBT as a Tracking Mechanism for 20 Year Bonds Index
TBT from the description on Yahoo: ProShares UltraShort 20+ Year Treasury (the Fund), formerly ProShares UltraShort Lehman 20+ Year Treasury, seeks daily investment results that correspond to twice (200%) the inverse (opposite) of the daily performance of the Barclays Capital 20+ Year U.S. Treasury Bond Index (the Index).
As a leveraged ETF, it suffers from decay in volatile market, but I wanted to see if I could find a model that would display the correlation between TBT and TLT (the index it is supposed to track). I cranked out ChaosHunter from Ward and used the daily close for 2010 and up to Friday. I used the first 200 days to create the model and the last 50 or so days for out of sample testing. Amazingly enough, the model tracked well in most of last year but shows now a much bigger divergence:
The out-of-sample period starts after the vertical dotted line. For the statistician, here are the numbers:
Either TBT is now a victim of decay or it is undervalued as compared to TLT. I will try post the results of a longer term model to see the effects of decay over multiple years.
As a leveraged ETF, it suffers from decay in volatile market, but I wanted to see if I could find a model that would display the correlation between TBT and TLT (the index it is supposed to track). I cranked out ChaosHunter from Ward and used the daily close for 2010 and up to Friday. I used the first 200 days to create the model and the last 50 or so days for out of sample testing. Amazingly enough, the model tracked well in most of last year but shows now a much bigger divergence:
The out-of-sample period starts after the vertical dotted line. For the statistician, here are the numbers:
Either TBT is now a victim of decay or it is undervalued as compared to TLT. I will try post the results of a longer term model to see the effects of decay over multiple years.
Subscribe to:
Posts (Atom)