Today’s title is inspired from the British Special Air Service (SAS) famous catchphrase, ‘Who Dares Wins’ – similar variations of which have been adopted by several elite army units around the world. The motto is often credited to the founder of the SAS, Sir David Stirling (although similar phrases can be traced back to ancient Rome – including ‘qui audet adipiscitur’, which is Latin for ‘who dares wins’). The motto was used to inspire and remind soldiers that to successfully accomplish difficult missions, one has to take risks (Geraghty, 1980).
In the world of economics and finance, the concept of risk is endemic to investments and to making decisions in an uncertain world. The ‘no free lunch’ principle in finance, for instance, asserts that it is not possible to achieve exceptional returns over the long term without accepting substantial risk (Schachermayer, 2008).
Undoubtedly, one of the riskiest investment instruments you can currently get your hands on is cryptocurrencies. The most well-known of them is Bitcoin (BTC), and its price has varied spectacularly over the past ten years – more than any other asset I have laid my eyes on in my lifetime.
The first published exchange rate of BTC against the US dollar dates back to 5 October 2009 and it shows $1 to be exchangeable for 1309.03 BTC. On 15 December 2017, 1 BTC was traded for $17,900. But then, a year later the exchange rate was down to just over $1 = $3,500. Now, if this is not volatility I don’t know what is!
In such a market, wouldn’t it be wonderful if you could somehow predict changes in market sentiment and volatility trends? In a hot-off-the press article, Shen et al (2019) assert that it may be possible to predict changes in trading volumes and realised volatility of BTC by using the number of BTC-related tweets as a measure of attention. The authors source Twitter data on Bitcoin from BitInfoCharts.com and tick data from Bitstamp, one of the most popular and liquid BTC exchanges, over the period 4/9/2014 to 31/8/2018.
According to the authors:
This measure of investor attention should be more informed than that of Google Trends and therefore may reflect the attention Bitcoin is receiving from more informed investors. We find that the volume of tweets are significant drivers of realised [price] volatility (RV) and trading volume, which is supported by linear and nonlinear Granger causality tests.
They find that, according to Granger causality tests, for the period from 4/9/2014 to 8/10/2017, past days’ tweeting activity influences (or at least forecasts) trading volume. While from 9/10/2017 to 31/8/2018, previous tweets are significant drivers/forecasters of not only trading volume but also realised price volatility.
And before you reach out for your smartphone, let me clarify that, although previous days’ tweets are found in this paper to be good predictors of realised price volatility and trading volume, they have no significant effect on the returns of Bitcoin.
- Does twitter predict Bitcoin?
Economic Letters, Dehua Shen, Andrew Urquhart and Pengfei Wanga (vol 174, pp. 118–22, Jan 2019)
- Who dares wins: The Story of the Special Air Service, 1950-1980
- The Notion of Arbitrage and Free Lunch in Mathematical Finance
- The Arbitrage Principle in Financial Economics
Arms and Armour Press, Tony Geraghty (1980)
Aspects of Mathematical Finance, Chapter 3, Walter Schachermayer (Springer 2008)
Journal of Economic Perspectives, Hal R. Varian (Vol. 1, No. 2, Fall 1987)
- Explain how the number of tweets can be used to gauge investors’ intentions and how it can be linked to changes in trading volume.
- Using Google Scholar, make a list of articles that have used Twitter and Google Trends to predict returns, volatility and trading volume in financial markets. Present and discuss your findings.
- Would you invest in Bitcoin? Why yes? Why no?