6. Empirical Results
6.2. Empirical Results Analysis II
The empirical results established by Analysis II complete the examination of Bi c in potential to serve as a safe haven by investigating the extent to which Bitcoin is fulfilling the liquidity requirement inherent in the definition of a safe haven (see section 1.2.). First, the implicit costs of trading Bitcoin are compared to other assets by means of the bid-ask spread to understand the degree to which Bitcoin can be bought or sold quickly at stable prices on a marketplace. Moreover, the bid-ask percentage
-50.00%
-40.00%
-30.00%
-20.00%
-10.00%
0.00%
10.00%
20.00%
30.00%
06/09/2019 13/09/2019 20/09/2019 27/09/2019 04/10/2019 11/10/2019 18/10/2019 25/10/2019 01/11/2019 08/11/2019 15/11/2019 22/11/2019 29/11/2019 06/12/2019 13/12/2019 20/12/2019 27/12/2019 03/01/2020 10/01/2020 17/01/2020 24/01/2020 31/01/2020 07/02/2020 14/02/2020 21/02/2020 28/02/2020 06/03/2020 13/03/2020 20/03/2020 27/03/2020 03/04/2020 10/04/2020 17/04/2020 24/04/2020 01/05/2020 08/05/2020 15/05/2020 22/05/2020 29/05/2020 05/06/2020 12/06/2020 19/06/2020 26/06/2020 03/07/2020 10/07/2020 17/07/2020 24/07/2020 31/07/2020 07/08/2020 14/08/2020 21/08/2020 28/08/2020
Weekly Returns: Bitcoin (INR) vs. S&P BSE 500
Bitcoin (INR) BSE500Index
84 spread of Bitcoin is briefly assessed against two financial stress indices to examine Bi c in liquidity development during times of market turmoil. Second, the explicit costs of trading are investigated by assessing the average transaction costs against the number of transactions for each specific day. Since investors flee to safe haven assets during crises, demand often rises, why it is important to know if the transaction fees increase when safe havens are needed the most.
6.2.1. Implicit Costs of Trading
Figure 7 and 8 depict the bid-ask percentage spreads of Bitcoin, gold, Apple, and Twitter in the period October 2013 through August 2020 (Figure 7) as well as for a more narrow and recent time frame from September 2019 through August 2020 (Figure 8).
Figure 7: Percentage Bid-Ask Spread 2013-2020 Bitcoin, Gold, Apple, and Twitter
Source: Bloomberg Professional Services (2020) and data.bitcoinity.org (2020)
It becomes apparent that Bi c in average bid-ask spread has generally declined, thereby signaling an improvement in Bi c in liquidity over time. From October 2013 to October 2016, Bitcoin reported the relatively highest spread across the considered assets reaching a peak of 12.14% in September 2014. Thereafter, Bi c in spread decreased and stabilized to values close to zero percent with gold beginning to report larger bid-ask spreads than Bitcoin. This image is supported by Figure
-5.00%
-3.00%
-1.00%
1.00%
3.00%
5.00%
7.00%
9.00%
11.00%
13.00%
01/10/2013 01/12/2013 01/02/2014 01/04/2014 01/06/2014 01/08/2014 01/10/2014 01/12/2014 01/02/2015 01/04/2015 01/06/2015 01/08/2015 01/10/2015 01/12/2015 01/02/2016 01/04/2016 01/06/2016 01/08/2016 01/10/2016 01/12/2016 01/02/2017 01/04/2017 01/06/2017 01/08/2017 01/10/2017 01/12/2017 01/02/2018 01/04/2018 01/06/2018 01/08/2018 01/10/2018 01/12/2018 01/02/2019 01/04/2019 01/06/2019 01/08/2019 01/10/2019 01/12/2019 01/02/2020 01/04/2020 01/06/2020 01/08/2020
% Bid-Ask Spread October 2013 to August 2020
Gold Apple Twitter Bitcoin
85 8, which demonstrates that Bi c in bid-ask spread was lower than gold, at a similar level to the volatile stock Twitter, but higher than Apple in the period from September 2019 through August 2020. Moreover, Figure 8 shows that the spreads of all the considered assets became more volatile and started reporting peaks as COVID-19 began spreading globally in March 2020. Thus, the liquidity of all four assets decreased at the outset of the COVID-19 crisis. This finding is supported by the three graphs in Appendix 9, which portray Bi c in bid-ask spread alongside the VIX and GFSI over time. These show that spikes in Bi c in bid-ask spread seem to move in lockstep with the sharp increases of the two stress indicators in March 2020.
Figure 8: Percentage Bid-Ask Spread 2019-2020 Bitcoin, Gold, Apple, and Twitter
Source: Bloomberg Professional Services (2020) and data.bitcoinity.org (2020)
To allow for a more precise comparison of the liquidity characteristics of Bitcoin, gold, Apple, and Twitter, Table 6 demonstrates the mean of each a e bid-ask spreads for 1) the entire sample ranging from October 2013 through August 2020, 2) a more recent sub-period ranging from September 2019 through August 2020, 3) a sub-period ranging from February 24th, 2020 to April 10th, 2020, thereby reflecting the period of high COVID-19 related market stress previously utilized in Analysis I. To test whether the differences in means between the assets during the three periods are significantly different from zero, this thesis refers to the results of the statistical significance test reported in Table 6.
0.00%
0.20%
0.40%
0.60%
0.80%
1.00%
1.20%
1.40%
02/09/2019 09/09/2019 16/09/2019 23/09/2019 30/09/2019 07/10/2019 14/10/2019 21/10/2019 28/10/2019 04/11/2019 11/11/2019 18/11/2019 25/11/2019 02/12/2019 09/12/2019 16/12/2019 23/12/2019 30/12/2019 06/01/2020 13/01/2020 20/01/2020 27/01/2020 03/02/2020 10/02/2020 17/02/2020 24/02/2020 02/03/2020 09/03/2020 16/03/2020 23/03/2020 30/03/2020 06/04/2020 13/04/2020 20/04/2020 27/04/2020 04/05/2020 11/05/2020 18/05/2020 25/05/2020 01/06/2020 08/06/2020 15/06/2020 22/06/2020 29/06/2020 06/07/2020 13/07/2020 20/07/2020 27/07/2020 03/08/2020 10/08/2020 17/08/2020 24/08/2020
% Bid-Ask Spread September 2019 - August 2020
Gold Apple Twitter Bitcoin
86 Table 6: Difference in Means Test Statistics
For the entire sample, Bi c in average bid-ask spread of 0.3787% is significantly higher than that of gold, Apple, and Twitter, thus indicating that Bitcoin has a low relative liquidity. As previously described and shown in Figure 7, this relatively high mean spread is, however, vastly influenced by Bi c in high spread in its early years from 2013 to 2016. Therefore, a look at the first sub-period provides a more current picture of Bi c in liquidity. The first sub-period shows that Bi c in mean spread of 0.0254% lies significantly below the bid-ask spread of gold and Twitter but above that of Apple. Hence, Bitcoin is relatively liquid when looking at a recent timeframe. When zooming in on the a e liquidity during the second sub-period, which reports high COVID-19 related market stress, it becomes apparent that Bi c in mean spread of 0.043% is higher than its mean of 0.02454%
during the first sub-period. This observation also holds for gold, Apple, and Twitter, which, in line with Figure 8 and Appendix 9, indicates that the liquidity of the four assets decreased during the period of high COVID-19 related market stress. Despite the decrease, Bi c in liquidity remains significantly better than that of the traditional safe haven gold. The results of the statistical significance test highlight that no significant inferences can be made about the difference in means between Bitcoin and Apple as well as Bitcoin and Twitter in the second sub-period. While Bi c in
Mean t-statistic p-value
Panel A: Entire sample period (01/10/2013 - 28/08/2020)
Bitcoin 0.3787%
Gold 0.0819% 13.6958 0.0000
Apple 0.0148% 17.1988 0.0000
Twitter 0.0198% 16.2480 0.0000
Panel B: Sub-period 1 (02/09/19 – 28/08/20)
Bitcoin 0.0254%
Gold 0.0932% -8.1969 0.0000
Apple 0.0173% 4.8123 0.0000
Twitter 0.0321% -6.0223 0.0000
Panel C: Sub-period 2 (24/02/20 – 10/04/20)
Bitcoin 0.0430%
Gold 0.2266% -4.0717 0.0003
Apple 0.0286% 1.5082 0.1365
Twitter 0.0358% 1.2650 0.2140
87 liquidity decreased under the high COVID-19 related financial market stress, Bitcoin appears to be more liquid than the traditional safe haven of gold and equally liquid as Apple and Twitter.
6.2.2. Explicit Costs of Trading
As outlined in section 2.1.1., every Bitcoin transaction must be added to the blockchain - the official public ledger of all Bitcoin transactions - in order for the transaction to be successfully completed and valid. Bitcoins cannot exist or be held independently of the blockchain. The validation of all transactions occurs through the process of mining, which takes care of including transactions in the limited space of a 1 MB block. When a block is filled up with transactions, it is added to the blockchain, which occurs circa every 10 minutes. Transaction fees are charged for this process, which make up the most substantial share of the overall fees charged when trading Bitcoins on exchanges.
While smaller, additional fees might be charged by the exchanges at which Bitcoins are bought and sold, this analysis solely focuses on the transaction costs of using the Bitcoin network and disregards the additional fees applied by exchanges, which differ across exchanges (CoinDesk, 2020a).
Table 7: Stylized Facts Transactions Fees
Entire sample period (01/10/2013
28/08/2020) Sub-period 1 (02/09/19
28/08/20) Sub-period 2 (24/02/20 10/04/20)
Mean 1.5246 1.4749 0.8094
Max 54.6380 6.4291 1.7856
Source: blockchain.com (2020)
Table 7 shows the mean and maximum of the average transaction fees per day during the same sub-periods, as outlined in the bid-ask spread section. While the mean of the average transaction fees for the entire period amounts to 1.5246 USD, the maximum observed average transaction fee totaled to 54.638 USD in December 2017. For sub-period 1, the average transaction fee was 1.4749 USD, with a maximum measured at 6.4291 USD at the end of July 2020. For sub-period 2, the mean and maximum of the average transaction fees amounted to 0.8094 and 1.7856 USD, respectively, thereby highlighting there to have been low transaction costs amid high COVID-19 related financial market stress.
88 Figure 9: Bitcoin Average Transaction Fee per Trade vs. Daily Transactions
Source: blockchain.com (2020) and charts.bitcoin.com (2020)
Figure 9 portrays the average transaction fee per trade per day and the corresponding number of transactions on that specific day in the period October 2013 through August 2020. Comparing the two suggests there to be a positive relationship between transaction demand and transaction costs.
For example, as the total number of daily transactions spiked in December 2017, so did the average transaction fee reaching the maximum value of 54.638 USD. While no later transactions reported similarly high fees, a relationship between increases in the number of transactions and the average fee can be observed at several points in time. During the ongoing COVID-19 period, the number of transactions, as well as average transaction fees, remained relatively stable with some spikes at the end of June as well as from the end of July to end of August. This suggests that Bitcoin was tradable at changing, but relatively low costs during the period with the highest COVID-19 related stress from February through April 2020, and at relatively higher costs during the summer months of 2020.