A Study of Bitcoin-Based Intraday Volatility Forecasting for Cross-Market Spreads

ABSTRACT

This study provides a volatility estimation based on cross-market spreads by analyzing the behavior of Bitcoin cross-market arbitrageurs. This study crawls real-time price data from different exchanges for empirical analysis and verifies the accuracy and validity of the method employed by comparing it with the existing mainstream methods. The following conclusions are drawn: 1) The more exchanges that can be utilized, the smaller the Bitcoin price volatility, and the larger the cross-market spread, the better the estimation effect of the proposed method; and 2) Volume had no significant effect on the estimation using our method.

KEYWORDS:

• Bitcoin
• volatility estimation
• cross-market spreads
• arbitrage

JEL CODES:

• G12
• G15
• G19

Acknowledgments

We would like to express our sincere gratitude to the editor and anonymous reviewers for their constructive comments and suggestions.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1. By the design of the Bitcoin network, it typically takes 10 minutes to transfer Bitcoins from one owner to another and confirm the transaction. The maximum time required to acquire Bitcoins from one exchange and sell them on another exchange is, then, 20 minutes.

2. Golden Finance: https://www.jinse.com/coin/Bitcoin. Investing: https://cn.investing.com/.

3. Each point in time includes raw minute data from multiple exchanges. After data collation, there are approximately two to four BTC/USD exchanges at each point in time. In addition, one to three Euro exchanges are included in the 2021 sampling data.

4. Likely due to the stricter regulation of blockchain currencies in China at the end of 2021; currently, Golden Finance no longer provides Bitcoin’s line information.

5. As mentioned earlier, the maximum number of minutes within a sampling time point is 300 or 120 minutes. The UC and CC tests give an overall statistic for a single sampling, so in adjusting the $\alpha$ values, this study first calculates the mean of volatility estimated by both methods within a single sampling $\overline{{\sigma }_1}$ and$\overline{{\sigma }_2}$, and then apply equations (18) and (19) for adjustment.

6. Volume information is not pushed by default on the Investing’s website; therefore, this article uses the average of the daily volumes of the two Bitcoin futures varieties BTC.CME and BTC00.CME within the Chicago Board of Trade, as a supplement to the 2022 volumes. Due to the different sources of transaction volume data in 2021 and 2022, we have displayed them separately.

7. Due to space limitations, only descriptive statistics of our CMSM method are provided here and in the subsequent grouping results.

Additional information

Funding

This work was supported by the National Social Science Foundation of China [18CJY060].