Can we anticipate the evolution of the bitcoin price by its seasonality? Is it a myth? In this paper, we will try to give an ideal overview of the comparative behavior of bitcoin since 2015. We will focus on monthly performances, effectively excluding shorter variations. The study of seasonality thus shows that the months of October, February and July are generally the most reliable and the most efficient. Will it be like this in the coming months?
A statistical overview
A first (unadjusted) measure of seasonality is to average the performance of each month of the year. Indeed, we then obtain an average monthly performance which can be sufficiently positive or negative to be interesting. The graph below, after data processing, summarizes the average monthly performance of bitcoin between 2015 and 2023.
Therefore, the most bullish months are (in order) October, December, February, November, July, April August and June. We find here this “Christmas effect” where the price of bitcoin tends to increase at the end of the year. Conversely, the most bearish months are on average January, September, May and March. Nevertheless, the movements seem a priori the most directional at the end of the year as well as at the peak of the beginning of the summer in July. The average performance of bitcoin is thus +24% in October and +20% in December. However, it was -6.2% in January and -4.8% in September.
Moreover, it is clear that this average monthly performance is not necessarily representative. Indeed, it is possible that the same month has a performance of +50% and another of -10% to establish an average of +20%. However, the probability of loss in our example is 50%. We therefore favor an adjusted seasonality of the bitcoin price.
Volatility-adjusted seasonality
To at least partially correct this variability effect, another measure is added. Indeed, for each month of the year, we measure the volatility of performances observed in the past. The lower the volatility, the greater the reliability of the observed average performance. That is to say, we would prefer a month that performs by +5% but with exceptional deviations from this performance, rather than an asset that performs by +25% on average but with past variations of -20%. .
The orange curve shows the level of volatility of all the variations recorded over the month in question. The black curve is consistent with the rods presented above. The months that interest us are therefore those with the lowest volatility and the greatest average performance. So, these are the months that perform the best and consist of sufficiently reliable movements. Thus, the months of January, May, August and November do not seem reliable enough to us. The volatility of these months actually exceeds 20%, even 25%.
The most reliable and successful months
To give a clear answer to our approach, we have represented an “index”. We divided the average monthly performance of bitcoin by its volatility. Thus, if the volatility of bitcoin over the month of September is very low (close to 0), then this will increase the value of the index. Therefore, the most interesting months are those with a very high index value (or very low in the negative). These will therefore be the months that best respond to the reliability/performance trade-off.
The most reliable and efficient months are therefore respectively October, February, July, and December, and September (on the downside). Conversely, the other months such as March, May, June, August, November and April do not seem reliable and efficient enough. We therefore observe that, in general, antagonistic and fairly directional movements take effect in September and October. Finally, it is interesting to note that in the case of adjusted seasonality, the “Christmas effect” is less noticeable.
Evolutions despite everything scattered
Once we have described the seasonality of bitcoin, we can better represent the different years comparatively. In fact, in July of each year, an index value of 100 is set. The years that ended in negative since 2015 as of July were therefore 2019-2020 and 2021-2022.
Nevertheless, we observe that the path of bitcoin diverges from one year to the next. The average of all these evolutions is represented in grey. We therefore observe that on average between 2015 and 2023, bitcoin tends to double between June of each year. The best performing year was 2016-2017, when the worst year was 2021-2022. Despite everything, we observe that the probability of major lows is higher in September and January.
The “ideal” trajectory of bitcoin
Therefore, it is possible for us to observe the average of these evolutions in a graph. The average of the indices presented above is isolated in the graph below. This graph shows the “ideal” evolution of bitcoin according to the months of the year based on the price history between 2015 and 2023. We see that the months of July, August and September are between them generally stagnant. A strongly bullish phase then generally begins from the end of September to December. Autumn therefore seems to be buoyant for the price of bitcoin. A new phase of stagnation begins between December and January, before a rebound in the following month.
But we have already specified that not all months have the same movement reliability. Despite everything, this trajectory makes it possible to describe a typical and ideal movement of bitcoin. We can therefore highlight a temporal effect on bitcoin. 62% of the time, the price of bitcoin goes up between (end) September and December. The average increase in the latter is more than 70% over the period in question. We therefore have a favorable seasonality in the fall for bitcoin.
Months with the lowest downside probability
We have seen that seasonality is a complex issue that does not stop at average performance. Another parameter to take into account is that of the frequency of declines over the 8 years studied. For example, if we observe 2 years with a negative month of January out of the 8 years, then the frequency of decline is 2/8 (25%). The study of the frequency of negative months thus shows the following conclusion…
The months with the lowest drop frequency are (in order) October, February and April. These are also fairly bullish months. Conversely, the months of September and August have a higher frequency of decline. September is therefore quite reliable on the downside (75% of Septembers are bearish), while the month of August is served by this observation. The other months have a drop frequency close to 50%. The October move is therefore not only low in volatility, it is also frequent and significant. The months of February and April seem to stand out in this sense.
A currently consistent bitcoin price?
The price of bitcoin rose sharply in January 2023 before stagnating in February and rebounding strongly in March. May 2023 was also a rather negative month. Also, while October 2022 was slightly bullish, November 2022 was significantly negative. If seasonality were a piece of information in its own right, the price of bitcoin could stagnate in the summer before picking up again at the end of the year. It is notable that the next halving will arrive in the spring of 2024 and that a bull run generally begins shortly before it.
In conclusion
In conclusion, the best performing months on average on the price of bitcoin are October, December, February, November, July. But this does not provide us with an indication of the reliability of this observation (the dispersion of past performance). The adjusted seasonality (of volatility) on the price of bitcoin shows that the most profitable months are October, February, July, and December, as well as September (down). Despite everything, the low number of years studied (8) and the persistence of a gap between the comparative evolutions prompt us to be cautious.
Thus we can refine our study with the frequency of declines over the months studied. In this regard, the months of October, February and April seem the most reliable with a high frequency of increases. A major observation of our study is therefore this effect ” autumn “… “62% of the time, the price of bitcoin goes up between (end) September and December. The average increase in the latter is more than 70% over the period in question”.
Receive a digest of news in the world of cryptocurrencies by subscribing to our new service of daily and weekly so you don’t miss any of the essential Tremplin.io!
