Cluster Chart Overview
As discussed in this article, cluster charts reveal relationships between positions within your portfolio or stocks within the broader market.
Related stocks tend to be directly connected and will form clusters (ie: groups) indicating that they tend to move together. On the other hand, positions that tend to behave independently within your portfolio will be connected to much fewer stocks and will typically not be a part of any clusters.
Since none of us know which positions will be profitable and which won't, our goal is to always maximize the number of independent and unique opportunities to be profitable throughout the portfolio. This is one of the concepts discussed as a part of this free video course.
If your portfolio has large and unbalanced clusters, you are running an inefficient portfolio that has unnecessary overlap and wasted capital. There are times where purposefully creating large clusters within your portfolio is a good strategy, however this is a more advanced topic that is discussed in the Diversify Portfolio members area.
Here we've got an 8 position portfolio:
When we move the cluster strength to roughly 50%, we see that there are 3 distinct groups within the portfolio.
The problem here is that the clusters are unbalanced and as a result represent:
- An inefficient portfolio
- 3 independent and unique opportunities to be profitable (ie: 3 positions disguised as 8)
- The larger cluster represents more risk to the portfolio in terms of the number of probabilistic outcomes (under the assumption that each position is equally weighted)
So where to from here? How can this portfolio be improved?
To improve the efficiency of this portfolio you need to go through each cluster and figure out which positions need to be removed so that they can be replaced with positions that provide a real and quantifiable benefit to you. There are several ways to do this which depend on factors such as your trading or investment style, fundamental vs technical analysis, risk reduction vs probabilistic outcomes, etc.
Ultimately how you decide which positions need to be replaced is up to you, with the goal being that each cluster is balanced. One way to do this would be to remove the most highly correlated positions from each cluster. For example here we have the correlation symbol rankings chart for this portfolio:
Looking at the companies from the biggest cluster, we can see that Facebook, Adobe and Google provide less diversification benefit to the portfolio than Amazon. Therefore one option might be to keep Amazon, and replace the other 3 companies with positions that provide more of a benefit to the portfolio in terms of diversification.
By going through each cluster and removing the positions which are ranked highest in terms of correlation (and therefore provide the least benefit in terms of diversification), we end up with a more efficient portfolio. You then have capital to spare which can be used to begin adding positions to the portfolio that provide a real and quantifiable benefit.
Of course using the above approach alone may not always be optimal. It will typically be a good idea to combine the above process with your own technical or fundamental analysis to determine which stocks stay, which stocks go and which new stocks take their place.
The cluster chart is a great way to see where possible inefficiencies in your portfolio lie. By combining that information with screeners and additional analysis, you ultimately end up with a more efficient stock portfolio.
Stay informed by joining our mailing list where you'll receive strategy, diversification and educational articles to help maximize your profit in the stock market while minimizing your risk.