Part 3: Pair/Arbitrage Trading Ideas
To take a brief departure from the indicator world, i would like to share with you all certain concepts—-some new, some old for pair trading and arbitrage. I will lengthen out this discussion to at least one other blog because there are so many ideas and nuances worth discussing.
A little creativity, thought, and common sense is just as useful as good indicators when figuring out how to go about extracting profit in “pairworld.” It is also important to think about this type of trading as much more of a high stakes poker game, where you have to know your opponents, and how they play in order to fully take advantage of all the opportunities that exist (more on this in later posts).
First, i want to categorize different strategies to make them easier to discuss. 1) Pure arbitrage: this doesn’t actually exist– even if you replicate the holdings of an etf continously and short or buy the ETF, there is no guarantee that you will realize a profit, as the market of buyers and sellers for the ETF are much larger than most individual traders and funds. 2) Near arbitrage: this is when we trade virtual substitutes–whether it is a leveraged ETF vs its unleveraged counterpart, or two nearly identical sector, style,or country ETFs. 3) Near substitutes: this is when there are distinct but slight differences between two sides of the pair—such as DBC vs GSG, where they are both commodity indexes with different weighting schemes that are still highly similar. Or it could be IYR–the real estate sector–vs VNQ– the REIT ETF. The other possibility is that you can synthetically create near substitutes (my favorite and very lucrative), such as trading the ten largest holdings in an ETF vs the ETF itself which may contain 30-50 stocks. This category also includes futures spreads such as USO (oil) vs USL (12-month oil) 4) Cross Pairs: this is where you have highly related pairs that are definitely different but are cointegrated by virtue of the fact that they are driven by similar macroeconomic factors–such as XLE–the energy sector vs USO, XAU the gold and silver equity index vs gold. In addition, cross pairs can also be spuriously related through a third party, such as the Canadian dollar- FXC vs the aussie dollar–FXA, which are both affected heavily by commodity prices.
Near arbitrage was one of the topics we briefly touched upon before in the form of leveraged ETFs. This is a topic i could write an entire book on, but just think of leveraged ETFs like Lego—you can really get creative and build immensely intricate and complicated strategies one brick at a time. There are 2x and 3x both long and short, as well as inverse ETFs covering identical sectors, industries or countries—sometimes even with multiple providers. As long as you can do some basic math, you are ready to get your hands dirty. The financial sector is a good example: there are 3 underlying unleveraged ETFs that are good to trade here vs their leveraged counterparts :IYF, XLF, and VFH. These can be traded against one another as a standalone strategy. In addition, you can short or go long these ETFs vs their 2x or 3x counterparts by simply using the requisite multiple (2x or 3x) of the dollar value purchased in the leveraged vehicle.
Before i get into this in the next post, lets introduce what i will call “basic strategy.” This is applicable to virtually all pair strategies without having to use any indicators at all. Basic strategy entails creating a baseline value measure of discrepancy which is scaled so that you can compare all of the different Lego blocks. As an example, you would look at the current % change for the day for say the XLF as your baseline, and then compare it appropriately to every leveraged counterpart, and near arbitrage or substitute. So if XLF is UP 2% today, we should expect UYG (2x long financials) to be up 4% (2×2—told you it was easy!). SKF (2x short financials) should be down 4%, VNF and IYF should be up 2%, FAZ (3x short financials) should be down 6%, FAS (3x short financials) should be up 6%, and SEF (the inverse or single short financials) should be down 2%.
From here the goal is to find the biggest “D” or discrepancy between the baseline and one of the related ETFs. “D” is basically a measure of how overvalued or undervalued on of the components is relative to the baseline. So in the above example, if SKF was down %5 instead of 4%, it would have a “D” of -1% or undervalued by 1%. The goal would be to get long SKF and balance it off with an offsetting long position–ideally something that is overvalued. This brings up tremendous opportunities to combine different positions to maximize profit. Squeezing out an extra .2% by taking an unconventional position can cover the fixed transaction costs making lower margin trades that would be passed on by the less saavy who believe them to be breakeven trades.
That’s all for now……..much more to come. Again please direct all questions to my email: firstname.lastname@example.org