The shortfall of ultra long and short

I like to apply idea to realistic situation other than on paper only. After all, you want to make real money profit, not paper profit.


I emphasized the use of ultra long and short ETFs. I want to make sure they are effective and how would they affect actual trading and profit in real world. Therefore, I did some very simple test using make up numbers.

Note that ultra long and short ETFs search for 200% exposure to a particular index on a day-to-day basis. Some articles emphasized the tricky part of day-to-day basis. However, this is not the important point in here, they went the wrong way.

The most important issue for the ultras are the leverage minus cost. What I mean is, when the index goes up, an ultra long ETF should produce roughly less than 2x, and when the index goes down, the ultra long ETF should lose roughly more than 2x. That is due to the cost and management fee.

I tried to produce a graph of a series of make-up returns. I initially assume the index loses 2% in 1 day, then gains 1% the next day. It flips between -2% and +1% for about 32 days. Then the pattern changes. It flips between +2% and -1% for about 70 days. The index produces approximately 19.4% in total. If we assume the ultra long produces exactly 2x exposure, the ultra produces approximately 39% at the end of 102 days.

Using the ultra actually compounded the effect. Therefore, if the index has been negative for a while, it takes more effort for the ultra long to catch up in performance. As you would see, while the index is slightly positive, the ultra long is still negative. But if the index continues to rise, eventually the ultra long picks up the speed and surpass the index. Conclusion, to hold an ultra ETF for long time, you have to make sure that you are making the correction directional bet. If the index takes longer than you expect to go to your direction, you have to expect to hold the ultra ETF longer to realize the compounding profit.

Now however if I change the exposure to 1.9x your direction and 2.1x the opposite direction, the picture totally changed.

If I use exactly the same index return pattern as above, assuming the ultra long only goes up by 1.9x when the index goes up, but the ultra long goes down by 2.1x when the index goes down to reflect the costs. Not very surprising, the ultra long actually underperformed the index by 0.02%.

Now if you believe this kind of exposure makes more sense than simply 2x. What you need to do is 2 things:

1. market timing matters. Just that you think the index should go up over a year, it doesn't mean you can earn more by using ultra long. Due to the imperfect 2x exposure, you better time the market carefully.

2. trade in and out. If there is no tax, sell into strength and buying back during pullback always help the performance of an ultra ETF. Given the compounding effect against you, I guess even after you consider the tax effect you will still find trading in and out helpful.