Hi, Dan. I am about 50% complete in backtesting 32 leveraged ETFs, an equal mix of both 2x and 3x funds, across various market indexes and sectors. I should have testing completed in about a week. I will post results both here and on InvestorsHub's AIMUSERS board. These are funds I have selected for my own use, but I am backtesting to identify issues and develop strategies with the AIM algorithm specific to these highly volatile funds. They're not everyone's cup of tea, and there are plenty of warnings on the internet about not holding leveraged ETFs long term, but I think those concerns are well mitigated with AIM managing the investments. The returns are almost always better, sometimes spectacularly better, than buy-and-hold. There are a few instances where buy-and-hold wins but it's not by a great deal, and the comfort of that cash cushion makes the AIM approach more appealing, in my mind, over being fully invested in the ETF to gain the maximum return.
I was initially doing backtesting with 10 years of historical price and dividend data in order to get a better feel for drawdowns for each of the ETFs, which are often quite volatile (which AIM loves). However, that was taking too long, so I scrapped that plan and started fresh with only 5-year backtesting. A couple of the ETFs are younger than that so I started on the fund inception date. I have identified a couple of websites that provide details on historical standard deviation (volatility) and maximum drawdown for each ETF, and the buy-and-hold returns with dividend reinvestment for the specific dates of my testing, for comparison. This eliminated the need for 10-year backtesting and allowed me to cut my testing time in half for each ETF.
One of the limitations of Lichello's algorithm is that cash can be depleted during significant drawdowns in share price, and this is particularly problematic for leveraged ETFs. I've learned some strategies from the AIM users on the board that helped, but was still kind of guessing. So I developed some drawdown curves specific to the settings the AIM model is using that helps determine the correct strategy to use, and that has helped a great deal.
My initial 10-year testing used a 50/50 equity to cash split, but I've learned to evaluate the 52-week price band and start my investment at the appropriate split based on where the starting share price falls in that band.
As I stated, I'm backtesting both 2x and 3x leveraged funds, including several you mentioned, like SSO, QLD, TQQQ and UPRO. I'm even testing a 2x Bitcoin futures ETF that performs extremely well, but it's only been around less than 2 years.
I look forward to sharing my testing results with AIM users. I'll also share my methodology for the initial equity ratio, AIM settings, and drawdown strategy used. I think that will answer a lot of your concerns using leveraged ETFs. |
