Come on, come on, let’s stick together!

Howdy folks, apologies for the awful song reference but I couldn’t think of a more appropriate title given this article will focus on how some stocks seem to act the same way, or stick together (in terms of price movements).  Now that I have made an average (at best) joke we can commence.

To begin, as always, we should ask why we’d care if certain stocks behaved in a way that was consistent (or not) with certain other stocks. You would rightly ask me why that matters when you can just read BLUELAKEINVEST and get behind Bellamy’s (ASX:BAL) and watch it go in only an upward direction. I don’t mean to alarm anyone but unfortunately good things don’t always last, and there needs to be a way to shelter your portfolio if one of your hero stocks takes a dive. You can Google search “Modern Portfolio Theory” if you want to know where this is heading, but basically heaps of finance geniuses have determined the best investment portfolio construction takes into account how the price changes in assets compare to each other and the market. Simple right? So basically, don’t have all dairy stocks in case there is an upcoming cow shortage and no milk can be produced. Own dairy and soy stocks so when your milk stock tanks your soy shares soar (so to speak) as presumably soy is used to make soy milk when it isn’t being turned into soy sauce or the like. That’s obviously an example, but the message is don’t hold just agriculture/mining (I’m so sorry)/finance etc.  industry assets.

So let’s take a squiz at how some BLI favourites perform (since May 2015, 150 trading periods).


So as you can see, Collection House (ASX:CLH) diverges from Bega (ASX:BGA), Neuren Pharmaceuticals (ASX:NEU) goes mental while ASX:QFN plugs along etc. It’s all very nice to look at, but how do you really really know which stocks will form a dream team? **MASSIVE DISCLAIMER, I graphed this backwards so most recent prices are at Time Period 1, do not be alarmed Bellamy’s is still doing very well. Just happened to notice as I was about to publish so my bad. Pls no hate.**

Luckily we have math nerds in the world who come up with things like Agglomerative Hierarchical Clustering (AHC, another one for the dinner parties yewwwwwwww) which in plain speak involves running an algorithm that looks at pairs of data in a set (like BAL and NEU prices on the same day) over a time period to determine which are most similar. Do not fear, there is a picture below. It’s called a dendogram, or tree diagram, and uses exactly the same data as the first graphs (in order this time).


So using the AHC model (noting that there are many many possible variants and this is only one) we see that the closest pairs are BAL and BGA (oh my god, we already predicted that here) and CLH and NEU. The hierarchy refers to how close the pairs are, with the best at the bottom of the tree. It’s worth noting that this is absolutely not how Modern Portfolio Theorists (MPT) come up with their portfolios, rather a more interesting way of getting at the same logic. If you read into MPT in your own time you’ll find it mainly sucks (i.e. crappy results when applied/studied using their formulas in the real world) so this is an attempt at using good reasoning (have a variety of stocks) with different maths.

TLDR; make sure that you have stocks that exhibit a wide range of behaviours and come from a lot of different industries so you’re safe if one of them crashes. Phew.