Why should we look at seasonality when observing the price of crypto?
Seasonality in mainstream market pricing is quite often an active area of inquiry for traders. The reason being that time-series repeatedly exhibit seasonality which consists of calculable patterns in data occurring at regular intervals.
The reasons for using seasonality to predict asset prices are these:
- Prices are determined by the decisions of people and the interaction that occurs when they buy or sell that asset, and
- Human behavior generally depends on the calendar (ie. vacation days, the work week, or other official holidays created by governments).
Despite the danger of data mining when looking for seasonal patterning, the reason it might exist in asset price is rationally based. Indeed, most macroeconomic gauges have a strong seasonal element and the government puts a lot of effort into developing models whereby data reports can be seasonally-adjusted.
In the equity markets, for example, researchers have identified calendar-related inconsistencies around the turn of the month and year. When we consider that digital asset markets are very reliant on the sentiment of people and rarely anchored to any particular fundamental value, it’s reasonable to probe further into whether or not there is some sort of seasonality happening.
How can we assess calendar-related anomalies?
Our team compiled a plot showing a calendar heat map of bitcoin’s four-week return over the last five years (Each square represents one day and green indicates that the past four weeks experienced a positive return, and red the opposite).
We chose a four-week return to smooth the data somewhat, ie. it’s easier to see what way prices are moving. The positive and negative indicator was chosen to deal with how bitcoin volatility has changed with time as it has matured into a legitimate asset class.
There are only a few years of reliable bitcoin price data, but some interesting observations immediately emerge:
Chinese New Year
In January 2018, the widely held narrative was that the Chinese New Yearwas responsible for the overall market sell-off and that prices would recover afterward. The story being that Chinese investors, withdrew capital from the crypto market for holiday expenses and gifts.
We wonder whether that is actually the sequence of events that is responsible for the market sell-off (for example, why doesn’t it exist during the run-up to Christmas?), but major market weakness has been observed in January and February for the past four years — roughly lining up with the Chinese New Year holiday.
Another narrative, of course, is that investors may defer selling their digital assets until after the first of January so capital gains on the sale won’t be applied until the subsequent year.
Annual capital gains payouts
Last March, another narrative was put forward as an explanation for the market sell-off: tax-related selling of crypto but this time to cover US capital gains taxes. The expectation was that the market would recover after the 15th of April, the imposed tax deadline.
For the last five years, bitcoin has averaged a 110% return per annum, generating a meaningful amount of capital gains for investors per year. Altcoins have seen even greater gains. Every year without fail, there has been a downturn in the market in late-March/ early-April, as some investors liquidate capital from crypto markets to pay their taxes.
This sell-off is not just seen in the crypto market — in April 2000, following the peak of the internet bubble, the NASDAQ Composite fell 9% in the week’s leak up to the tax cutoff.
The fall trading season
The last three months of the year has historically shown to be where price gains in crypto are more likely. Both this bubble peak in the price of bitcoin and the previous peak in 2013 happened in December. However, the price history we have for bitcoin is limited, and we have yet to find a compelling narrative for why gains are more likely in the fourth quarter.
There is a greater internet usage during the fall and winter months. And, as mentioned above, investors may delay selling until January to defer taxes. But neither of these explanations seem very compelling, and this may just be a non-recurrent coincidence.
Whether the sequence of events is true or not, if enough investors adopt a narrative and allow it to alter their trading decisions, seasonal narratives have the ability to be self-fulfilling. This year, the tax deadline narrative was so widely shared and enough people acted on it that after the tax deadline, bitcoin briefly rallied from $7,000 to $10,000 USD.