A first take at measuring Bitcoin Volatility

Bitcoin has seen incredible attention from commentators and the general public following spectacular price moves over the previous 12 months. In almost any discussion regarding Bitcoin as viable currency, attention inevitably turns towards its undeniably high volatility. This volatility, in contrast to more widely traded currency pairs, is often provided as evidence that Bitcoin is not and cannot be a currency. Such claims are often unsubstantiated and require closer inspection and data. Our preliminary analysis using major USD Bitcoin exchanges, shows a slight but statistically significant decline in volatility since USD trading began.

Bitcoin Volatility

Classic monetary theory tells us that there are three main uses for money: a medium of exchange, a store of value and a unit of account. While all of these different uses require different attributes of the media used as money, a stable value is essential for the normal functioning of an economy. Indeed as the recent price fall has reminded investors, asset prices rise and fall and hence with high volatility it might not serve as a good store of wealth and even worse as a currency.

The price discovery of Bitcoin has been an interesting journey. There is always wide speculation as to why the Bitcoin price is moving in either direction. News and media hype still probably play a large role but there is some evidence that the volatility is slowly subsiding albeit from a relatively high starting point.

In our analysis we use used data from Bitstamp, MtGox and BTC-e. We analyzed daily price data using a GARCH(1,1) model. We then took the average of the volatility estimates. There are other classes of models that we are looking at to capture the underlying data better. In our next post we will look at the distribution of returns and look at what this means for what model of volatility should be used.

On average Bitcoin is ten times more volatile than the S&P 500 or 15 times more volatile than the EUR:USD corridor. The S&P 500 has a daily volatility of approximately 0.5% and the EUR:USD corridor has a daily volatility of approximately 0.35%. Bitcoin’s daily volatility on average is 5%. Currencies typically have much more stable values to fulfill their purposes demanded by their users. Even the most volatile currency pair ZAR:JPY has a daily volatility of close to 1%.

Bitcoin is however unlike normal fiat currencies. Its inner workings, while transparent, are not well understood and it does not have sophisticated markets to determine its price. It is therefore unsurprising that the road to price discovery is rocky and bumpy. There is plenty instances of asymmetric information as large holders can be owners of Bitcoin businesses and have inside knowledge of the latest regulatory news and other market sensitive information. There are limited routes for people who are willing to short the market to gain exposure limiting the channels of information transmission. Although we remain skeptical about the ability of futures markets to resolve the issue of volatility, Bitcoin may benefit from some of these channels being opened and might lead to a stabilizing effect.

Despite the high volatility of Bitcoin, there are also prolonged periods of lulls in volatility. There are many possible interpretations to such a distribution but it seems like the most conclusive evidence that the price of Bitcoin is informed by new news stories breaking. If all the participants in the market were well informed about the current state of Bitcoin and its course in the near future, the price would not react as much to new news stories. However, large spikes in volatility indicate that the market is reevaluating the price to a large degree on the announcement of new news. The recent price fall on the announcement of new regulation in China is a prime example.

Trend in Volatility

A simple linear probability model shows a decline in the expected volatility of any given day over the past four years from around 9% to approximately 4.5% today. As seen from the chart above, a linear model cannot capture the structure of the data very well but does provide a first look at the trends in volatility that we are observing. As the volatility of the currency remains high we expect many merchants that accept Bitcoin to rapidly convert it back into fiat currencies to avoid the currency risk. Regulators and investors are right to pay close attention to the volatility of Bitcoin and we look forward to continuing our analysis and giving clear measures of the Bitcoin economy.

Postscript: Below are the measures of Volatility across the different exchanges across the different time series: Opening price, High price, Low price and Closing price

Complete Data All Exchanges


2 thoughts on “A first take at measuring Bitcoin Volatility

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s