Author: Eyal Beigman, Gerard Brennan, Sheng-Feng (Philip) Hsieh, and Alexander Sannella.1
The statements in this document reflect guidance issued as of May 4, 2020.
This article briefly discusses a methodology for fair value measurement of cryptocurrency pairs which are thinly traded or not directly traded on exchanges. The translation, from Binance Coin (BNB) to U.S. dollars (BNB-USD), will be used as an example to illustrate the outcome of the methodology.
Lukka Prime2, a leading cryptocurrency pricing product that leverages a mark to market methodology to provide reasonable fair value, measures for actively traded cryptocurrency pairs. Following accounting guidance of fair value measurement from ASC 820 (FASB) and IFRS 13 (IASB), Lukka Prime dynamically determines the principal market for each cryptocurrency pair by considering static characteristics, short-term and long-term trade behaviors among exchanges worldwide, and treats the last quoted price of the currency pair on the determined principal market as the fair value measure (level 1 input per the ASC 820 guidance) at that specific moment. To obtain reasonable fair value measures for thinly traded currency pairs, however, a mark to model approach is needed because of the limited number of observable transactions. As such, Lukka Prime may not be the most appropriate valuation method for thinly traded pairs.
The Valuation Methodology for Thinly Traded Cryptocurrencies Suppose that one needs to translate cryptocurrency ABC to the U.S. dollar for financial reporting purposes; it is intuitive to use the exchange rate of ABC-USD directly as reference. However, prior literature has indicated and evidenced that the transaction volume on financial markets, including cryptocurrency ones, is associated with the market credibility (Nasiri, Bektas, and Jafari 20183) and works as a channel to incorporate private and public market information (Bianchi and Dickerson 20204; Brandvold; Molnár, Vagstad, and Valstad 20155; Makarov and Schoar 20196; Park and Chai 20207; Sockin and Xiong 20208). In other words, the information embedded in the exchange with higher transaction volume for each specific currency pair might be more credible.
One could obtain two more alternative exchange rates for ABC-USD with the intermediate currencies DEF and GHI, rather than the direct exchange rate ABC-USD, by multiplying the exchange rate of ABC-DEF and DEF-USD or ABC-GHI and GHI-USD. Supported by the prior literature, it is proposed that the translation using exchange rates from cryptocurrency markets with higher transaction volume would be more representative and reliable because of market credibility and the incorporation of public and private market information. This is the theoretical foundation of the methodology.
Figure 1: Different paths that translate cryptocurrency ABC to the U.S. dollar (USD)9
The valuation methodology for thinly traded cryptocurrencies starts with listing all possible paths from the target cryptocurrency (currency ABC) to USD. The number of pairs comprising a path should be equal or higher than two. The exchange rate for each pair in a path is estimated and obtained from Lukka Prime10. The determined optimal and representative path translating the target cryptocurrency to USD is the path with the highest “bottleneck volume” from all path candidates.
For both alternative paths from ABC to USD, ABC-> DEF->USD and ABC->GHI->USD, the pairs with the minimum transaction volume, or the “bottleneck volume,” in both paths are DEF-USD and ABC-GHI, respectively11. The higher the transaction volume, the more credible the market and its exchange rate would be. Hence, the selected optimal path in the valuation methodology is the path with the higher bottleneck volume in both path candidates. In the example, because the higher bottleneck volume between DEF-USD and ABC-GHI happens on the DEF-USD leg, the determined optimal path to translate ABC to USD is ABC->DEF->USD.
Pilot Test Results
Binance Coin (BNB) is one of the most popular cryptocurrencies worldwide with the eighth highest market capitalization12. However, BNB could not be directly traded to USD on major exchanges13. It is a need to dynamically obtain an objective and representative fair value measure for BNB-USD for financial reporting if companies are holding BNB.
In the pilot test, BTC, USDT, and ETH are selected as intermediate currencies between BNB and USD to create possible paths for the illustration of the methodology. Figure 2 presents four candidate paths to translate from BNB to USD, including BNB->BTC->USD (path 1), BNB->ETH->USD (path 2), BNB->USDT->BTC->USD
(path 3), and BNB->USDT->ETH->USD (path 4). The period of cryptocurrency transaction data covers from August 1 to December 31, 2018. The most optimal and representative path is analyzed and determined on a minute-by-minute basis14.
Figure 2: Different paths that translate Binance Coin (BNB) to U.S. dollars (USD) (Relative transaction volume is not shown in the figure.)
Table 1 summarizes the distribution of each path candidate determined as the most optimal and representative path from BNB to USD over the sample period. Path 1 (BNB->BTC->USD) dominates and occupies 83.29% of the sample period to be determined as the optimal path under the methodology; 11.28% of the period identified the optimal path as Path 2, BNB->ETH->USD; and 5.43% of the period identified the optimal path as Path 3 (BNB->USDT->BTC->USD). Surprisingly, Path 4, BNB->USDT->ETH->USD, is not selected to be the path. This means that the bottleneck volume in Path 4 is not the maximum bottleneck volume among four path candidates in the sample period.
Table 1: The ratio of the determination as the optimal path for the translation of BNB-USD (The sample period: August 1 to December 31, 2018)
Figure 3: The ratio of the difference between the maximum and the minimum of BNB-USD exchange rates from the four candidate paths
To prepare financial reporting, managers may determine the exchange rate to translate from BNB to USD by subjectively selecting one of many alternative paths. However, using a method that does not follow an objective approach, such as the one outlined in this article, reduces the reliability of financial reporting for thinly traded cryptocurrencies.
Figure 3 presents the ratio of the difference between the maximum and the minimum of BNB-USD exchange rates from the four path candidates from August 1 to December 31, 2018. Although the average difference is only 0.005 (0.5%) for the whole sample period, it fluctuates in some specific time intervals. As shown, for instance, the peak appeared in the five-day period, October 14 to 18, 2020, and the average difference reached 0.0157 (1.57%), with a peak of around 12%. The cryptocurrency being valued, the temporary changes in market condition and investors’ trade behaviors, and/or the time period in which the fair value is estimated are potential reasons contributing to the relatively large fluctuation.
Built based on the theory that higher transaction volume is linked with higher market credibility, the methodology considers transaction volume and aims to provide a more representative and credible fair value measure for thinly traded cryptocurrencies. The methodology specifically would be appropriate to value thinly traded cryptocurrencies as well as ones not directly traded to U.S. dollars on exchanges. Aligning with current fair value measurement guidance under ASC 820 and IFRS 13, the methodology could be executed automatically and dynamically to provide valuation for financial reporting.
1 The statements in this document should not be treated as legal, tax, or accounting advice. The document is intended to provide general information only. If a person would like such advice, they should seek professional advice with regard to their specific facts. The statements in this document reflect guidance issued as of May 4, 2020.
2 More details and information about the Lukka Prime could be reached from the Lukka Prime pricing webpage. Available at: https://lukka.tech/lukka-prime-pricing/
3 Nasiri, S., E. Bektas, and G. R. Jafari. 2018. The impact of trading volume on the stock market credibility: Bohmian quantum potential approach. Physica A: Statistical Mechanics and its Applications. 512: 1104-1112.
4 Bianchi, D., and A. Dickerson. 2020. Trading volume in cryptocurrency markets. Working Paper. Available at: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3239670
5 Brandvold, M., P. Molnár, K. Vagstad, and O. C. A. Valstad. 2015, Price discovery on Bitcoin exchanges. Journal of International Financial Markets, Institutions and Money, 36: 18-35.
6 Makarov, I., and A. Schoar. 2019. Price discovery in cryptocurrency markets. AEA Papers and Proceedings 2019, 109: 97-99.
7 Park, M., and S. Chai. 2020. The effect of information asymmetry on investment behavior in cryptocurrency market. Proceedings of the 53rd Hawaii International Conference on System Sciences. Available at: https://scholarspace.manoa.hawaii.edu/handle/10125/64236
8 Sockin, M., and W. Xiong. 2020. A model of cryptocurrencies. Working Paper. Available at: https://www-nberorg.proxy.libraries.rutgers.edu/papers/w26816
9 The thickness of arrows represents the relative transaction volume of each currency pair.
10 Among exchanges could trade the specific currency pair, Lukka Prime is able to follow the accounting standards (ASC 820 and IFRS 13) on the fair value measurement and dynamically determine the principal market and the fair value of the specific currency pair at any given moment.
11 Observed from the thickness of arrows in Figure 1.
12 The information about the market capitalization and the following transaction volume percentage for each cryptocurrency pair on individual exchanges is derived from CoinMarketCap (https://coinmarketcap.com/) at 3:00 PM, May 4, 2020. The percentage is dynamic and time-variant, depending on when you derived the information.
13 There are currently 10 reliable exchanges, mentioned in the Bitwise’s presentation to the SEC (Bitwise 2019). Available at: https://www.sec.gov/comments/sr-nysearca-2019-01/srnysearca201901-5164833-183434.pdf
14 Different time intervals, including 5-minute, 10-minute, and 60-minute windows, are also analyzed and indicated similar results (untabulated).