Cryptocurrency Relationships Revealed — (Correlation Heatmaps)
Have you ever noticed that when the Bitcoin price goes up, some Altcoins also go up? In the same way, when the Bitcoin price goes down, other Altcoins also gain value?
This is proving to be quite a common pattern in the cryptocurrency world, and that doesn’t just go for Bitcoin. Lots of coins are correlated with each other so it’s a good idea to measure their movements in relation to each other. A good portfolio strategy will consider how coins are intertwined, and how their relationships affect each other. So I decided to investigate this using the Pearson Correlation Coefficient.
The Pearson Correlation Coefficient
I started off by gathering cryptocurrency price data over the years to analyse the correlation coefficient, then I will see how correlation changed over the years. I used a Python script with Jupyter notebook to carry out the analysis.
But what is the correlation coefficient? For this study, I used the Pearson Correlation Coefficient. This according to Wikipedia is:
“a measure of the linear correlation between two variables X and Y. It has a value between +1 and −1, where 1 is total positive linear correlation, 0 is no linear correlation, and −1 is total negative linear correlation.”
With the Pearson Correlation Coefficient, we can get any number for a currency pair between -1 and 1, which will tell us how correlated the currency pair is.
But what does this mean for cryptocurrencies?
● A correlation coefficient of 1 means the two currency pairs are perfectly correlated and will move in the same direction 100% of the time, they have a perfectly positive correlation.
● A correlation coefficient of -1 means the two currency pairs will move in the opposite direction 100% of the time, they have a perfectly negative correlation.
● If the correlation coefficient is 0, it means that the currency pairs have no correlation relationship.
Advantages
● To avoid investing in two coins that cancel each other out:
Imagine you invest in two different cryptos but the correlation coefficient tends to -1. This means as one is going up, the other will go down, thus cancelling out the value of your investment.
● Hedging:
Similar to the first point, but if it’s part of your strategy to hedge a certain investment, you could choose a crypto which would move in the opposite direction of your initial investment to make up any losses in case the first coin loses value.
● Diversify the Portfolio
It could be that you have a crypto that you invest in, but instead of going in 100% on it, you could find another closely correlated coin (let’s say a coefficient of 0.7) and split some of the investment onto this coin to diversify the portfolio but keeping the same momentum.
Disadvantages
● Correlations of currency pairs will change over time (we’ll come to this in a minute).
● This type of analysis can only be done for linear dependencies between two assets. But it’s pretty unusual to have perfect linear dependencies between different assets.
TL;DR Using the correlation is a good way to have a quick overview of how coins are related to each other. However, this type of analysis should be used alongside other research to get an overall, accurate picture. Correlation analysis can help with strategic portfolio decisions.
Correlation Heatmaps
I carried out a correlation analysis between the following currencies:
- BTC
- XRP
- XMR
- XEM
- STR
- LTC
- ETH
- ETC
- DASH
- BCH
The analysis was done for the years 2016–2018. This is to see how the relationship between coins changed over time.
This correlation heat map is from the year 2016. The diagonal is perfectly correlated, as expected, because coins are paired with themselves, so we can ignore this. We can also see that there is not data for BCH which again is expected. The thing that stands out here is the correlation between BTC/LTC.
We also see a fairly positive correlation between XRP/STR. This is expected with these coins because Bitcoin and Litecoin are part of the same family, just like Ripple and Stellar.
This correlation heat map shows us the relationships in 2017. Now this is interesting as we are seeing the majority of coin pairs starting to align in positive correlation.
Finally the 2018 correlation heat map. All the coins here are even more positively correlated with each other. We can see the Pearson correlation coefficient over 0.5 for nearly all the coin pairs.
What is extremely interesting here, is that every year, the different cryptos in the market are becoming more positively correlated with each other. And it isn’t just pairs of cryptocurrencies from the same family but all cryptos across the market.
In my previous article Is it Time for a Bitcoin Come Back? I talked about the Bitcoin/Altcoin cycle. If Bitcoin moved in one way, the Altcoins would move in the other, following the Bitcoin Domination pattern. This still stands as I explained it, but we can now also consider that with this correlation analysis to determine how strong this cycle will be in the future.
Conclusion
It’s interesting to see how cryptocurrencies correlate in the market. At first they were quite independent from each other. Over a very short time, this has changed and we can see an overall positive correlation pattern for all of the crypto pairs. Using this as a gauging tool along with other research gives dimension to the overall bigger picture to get a head start view on how the crypto markets are changing.
