External Relative Strength

I’ve been doing some research around something called External Relative Strength (ERS). ERS measures the stock’s price performance relative to all other listed equities. Basically, it measures how well or poorly a stock is performing relative to its’ peers. Our research suggests that the market has a strong propensity to mean revert in the short-term, in other words, recent underperformers, as measured by ERS, should on average outperform and vice versa. With this in mind I built a simple strategy that attempts to exploit this tendency. I created a simple 2 day ERS that compares the two day rate of change in price for each stock with all other listed stocks. The ERS measures the percentage of stocks that currently have a two day rate of change in price reading below the stock in question. My rules are as follows:

Entry: ERS < 5 (the stock is underperforming 95% of the universe on a two day basis.)

Exit: ERS > 20 (stock is outperforming 20% of the universe)

The results were surprisingly good for such a simple strategy. Before we analyse the performance results, let’s have a quick look at ERS applied to a chart. To refresh your memory, I’m using a 2 day ERS that measures the cross-sectional relative strength of each stock listed on the JSE. The chart below will help clarify ERS. High readings in the indicator imply that the stock is outperforming the majority of listed equities, while low readings tell us that the stock us underperforming.

Due to the markets powerful tendency to mean revert, one approach would be to buy stocks that are underperforming the market in the short-term (low ERS reading) and exit them when their performance improves relative to the market, such as the study we’ll discuss today. But before we do, here are a couple of other interesting ideas that employ ERS:

1) One could also develop sophisticated pairs trading methodologies that hold underperformers long and outperformers short, achieving rand neutrality in the market I.e. equal amounts of rand exposure on both the long and short side, or perfectly hedged. A purely neutral strategy such as this has the desirable property of being self-financing in that the proceeds from short holdings can be used for long holdings, effectively requiring zero investment to run.

2) Taking this a step further, complex position sizing algorithms could be used to allocate capital based on the stock’s ERS reading. For instance, the best out/underperforming stocks would receive greater allocations relative to other positions because they offer the highest probability to profit. Of course, the permutations are endless, and that’s precisely why I’m so passionate about the markets: they’re analogous to a complex puzzle with infinite solutions across differing degrees of effectiveness.

Today we’re going to review the performance statistics of a simple strategy that employs the following two rules:

Entry: ERS < 5 (the stock is underperforming 95% of the universe on a two day basis.)

Exit: ERS > 20 (stock is outperforming 20% of the universe)

And here are the performance summary statistics net of broker fees at 0.2%:

Clearly a lot of work would still needs to be done with this model. Returns could be dramatically improved by using a simple trend filter, volatility position sizing algorithm or Top 40 constituent filer (only trade stocks listed in the Top 40 Index) among others. However, strategies that demonstrate potential in their raw form is the starting point of all successful system design, and there’s enough evidence here to suggest further investigation to be worthwhile.

Happy Trading,
PJ

Passionate about generating and sharing quantified trading models that empower individuals to trade successfully. I founded www.sutherlandresearch.com to realise my passion.

Leave a reply:

Site Footer