Risk measurement of Bitcoin pricing

Abstract: This article discusses the risk measurement of Bitcoin pricing, and briefly introduces its application in high-frequency quantitative strategies and futures margin setting. A deeper understanding of the Bitcoin market may provide more reference for regulators.

There is a spread in the price of Bitcoin on various spot exchanges. In order to represent the market consensus price of Bitcoin, Bitcoin Derivatives Exchange uses its own custom index, such as BitMEX’s .BXBT index (calculated based on the weighted average of the “latest transaction prices” of each component exchange used by it ) And Deribit’s BTC- USD index (calculated based on the average value of the “best bid and best asking price” of each component exchange used by it).

The market consensus price of Bitcoin can be regarded as the price of Bitcoin. Due to the spread of bitcoin prices on various spot exchanges, the specific value of this pricing depends on the calculation method we use, and different spreads should have different effects on the accuracy of pricing. For example, although in most cases arbitrage robots will make the bitcoin prices of various spot exchanges converge, sometimes the main pull will make the prices of some exchanges much higher than other exchanges. These different situations will make The accuracy of pricing is affected, so we introduce the risk measurement of Bitcoin pricing. Using this as an indicator can guide application scenarios such as high-frequency quantification strategies.

Although the high-frequency quantitative strategy is based on the current Bitcoin price to perform corresponding operations, due to a certain delay between the order and the exchange, the profit and loss of the strategy is actually determined by the Bitcoin price in the next instant. The current reasonable pricing of Bitcoin and the degree of dispersion between the prices of various spot exchanges will directly affect the price of Bitcoin in the next moment. Therefore, it is meaningful to introduce the risk measurement of Bitcoin pricing to measure this degree of dispersion.

So, what kind of data is used to measure the risk of Bitcoin pricing? We believe that the current buy 1 price and sell 1 price of the mainstream spot exchanges should be used, and only current data can affect the bitcoin price in the next moment. The detailed description is given below.

What kind of data is used for analysis

1. Use rate of return data?

Do you use yield data for analysis? Literature [1] pointed out the reasons for adopting rate of return data:

Most financial research focuses on asset returns rather than asset prices. Campbell, Lo, and MacKinlay (1997) give two main reasons for using returns: First, for ordinary investors, return is an investment opportunity A complete and free-scale overview; second, the rate of return series is easier to handle than the price series and has better statistical characteristics. Commonly used rates of return include single-period return, multi-period return, continuous compound return, asset portfolio return, and excess return.

However, the calculation of the rate of return depends on the price of Bitcoin. The future price of Bitcoin is difficult to predict. Its past price can be used to calculate the historical rate of return. However, it is difficult for these rate of return data to directly reveal the future price trend, so the return is not used here. Rate data are discussed.

2. Use historical prices?

In the future, the price of Bitcoin may go out of a market similar to that of history, but in this article, we do not believe that history will repeat itself, so we do not use its historical price data for discussion.

Nevertheless, historical prices are still very valuable. For example, using historical price and trading volume data to analyze the situation of the main force collecting chips, washing and testing, and distributing chips. This has a certain guiding significance for the understanding of the subsequent market, but this is the content of another article.

3. Use real-time transaction price data?

When there is an extreme market situation and no market order is executed, it is possible that the market maker’s quoted price on the market may deviate too much from the last real-time transaction price. In this way, the real-time transaction price does not accurately reflect the price of Bitcoin at all times, so real-time transaction price data is not used here.

4. Use the data of buy 1 price and sell 1 price?

Since the buy 1 price and the sell 1 price can reflect the accurate price of Bitcoin in time, this article chooses to use them for discussion.

Here is a brief introduction to Markov chain theory:

There is a sequence of states in the system, and different moments correspond to a state of the system, and k is any moment. Markov chain refers to a Markov random process with Markov property, and Markov property is also called no aftereffect. No aftereffect indicates the state of the system at time k+1 in the future, which only depends on the state at the current time k, and has nothing to do with the state at any previous time.

Therefore, this article actually believes that currency prices have Markov characteristics. That is, we use the current buy 1 price and sell 1 price data for analysis, and think that the current data can have an impact on the market in the next moment, and the historical data has nothing to do with the market in the next moment (in other words, it has nothing to do with the market in the next moment. Only the current various data are really relevant to the market, they can really exert an influence on the next moment, and except for coincidences, history cannot repeat itself in the next moment).

This article is not to deny the possibility that some cyclical factors revealed by historical laws have an impact on currency prices. However, after analyzing some basic conditions, and then adding to the discussion of historical laws, it may be possible to thoroughly discuss the issue.

Further discussion on the data of buy 1 price and sell 1 price

1. Include the buy 1 price and the sell 1 price of the same exchange into the analysis at the same time

For the same exchange, both the buy 1 price and the sell 1 price have an impact on the pricing of Bitcoin. When the two divergence is large, there will be a large price difference; when the two divergence is small, the price is very close. Therefore, the two are included in the analysis at the same time, so that the selected data can also reflect the information on the price divergence of the same exchange.

2. Incorporate the buy 1 price and sell 1 price of different exchanges into the analysis

The price of Bitcoin on all major exchanges is different, and there is a certain spread. Regardless of the price difference, the buy 1 price and the sell 1 price of different exchanges have an impact on the pricing of Bitcoin, so the buy 1 price and the sell 1 price of different exchanges should be included in the analysis.

And if you have to point out which exchange’s price is the most reasonable, you can think about it like this: Under normal circumstances, it seems that the range with a high degree of price concentration is more reasonable; but sometimes the opposite is true. It is more reasonable to be far away from the concentration range, for example , The main force pulls orders on a certain exchange, which makes the bitcoin price of this exchange greatly deviate from the price of other exchanges. At this time, perhaps it should be more inclined to recognize the main force’s perception of the price. This price is more in line with the basis of reasonable pricing judgment .

Risk measurement

tzHCyL8VSqgrOq5t0kOITPNLZr97oEJ36ub85r2l.png

2. Risk measurement of Bitcoin pricing

The data we use to buy 1 price and sell 1 price comes from the following exchanges: Binance, Bitstamp, Coinbase Pro, FTX, Gemini, Huobi, Kraken, OKEx. The following is the data at 2021.8.17 11:47:25:

7wUl69D1NnZOLaOAArdm30VYJnpHCpj4um8L7zkd.jpeg

Figure 1

In the above figure, the red one is the selling price of each exchange, which has been arranged in descending order for the convenience of observation; the green one is the buying price of each exchange, which has also been arranged in descending order. Some buy 1 price is higher than sell 1 price, indicating that there are arbitrage opportunities between different exchanges.

liTWhhOLc8qnDvqJLooWLl1yKGq9gV01AVGhSFtH.png

Picture 2

According to the data of buying 1 price and selling 1 price of each exchange at a certain time, the above figure shows the variance of these data, which indicates the risk of Bitcoin pricing. The figure above also shows the historical variance of the previous period.

Application

1. Application in high-frequency quantitative strategy

The greater the variance (risk), the more unstable the current Bitcoin pricing. The next moment (also the moment we place the order at the exchange), the degree of price fluctuations will increase, which may make the original profitable The operation becomes unprofitable.

Therefore, this variance (risk) can be used as an auxiliary indicator to participate in the risk control of high-frequency strategies. For example, when the value of this indicator is large, our strategy does not operate.

In addition, compared with historical data, when this variance is particularly large, it indicates that there is a main force pulling or smashing in some exchanges, which may indicate that the extreme market is about to come-but this requires more testing. And verification. In this case, the defensive measures corresponding to the strategy should be activated immediately.

2. Futures margin [2]-[3]

At present, the margin system in the futures market is mainly divided into a static margin system and a dynamic margin system.

Under the static margin system, the initial margin and transaction maintenance margin are fixed. Under the dynamic margin system, the ratio of margin will be dynamically adjusted with the fluctuation of the futures contract price. The advantages of this dynamic margin collection method are obvious. It has a correlation with the contract price, so it can well capture the Changes in market risks caused by futures price fluctuations can better compensate for the risks caused by contract price fluctuations. The static margin system cannot meet this requirement. Therefore, academia generally believes that the dynamic margin system is an inevitable trend in the development of the futures market.

The maintenance margin for futures trading should not be too high or too low: If the maintenance margin for futures trading is too high, although the possibility of default will be reduced, considering that the maintenance margin for futures trading is an important part of transaction costs, excessive futures trading Maintaining margin will suppress investors’ enthusiasm for investment, and the market’s liquidity will also be drastically reduced. In the long run, it is not conducive to the sustainable development of the futures market. On the contrary, if the maintenance margin for futures trading is too low, the probability of default will be greater, which is not conducive to the stability of the futures market and will ultimately harm the interests of investors. Therefore, the maintenance margin for futures trading should be set reasonably.

A large amount of literature links the setting of maintenance margin for futures trading with the VaR risk measurement method. This method uses mathematical statistics to determine the futures trading margin, which has strong scientificity and shows good applicability, which can better make up for many deficiencies in economic models. However, the accuracy of this method largely depends on the risk measurement accuracy of the VaR method. Studies have shown that the shortcomings of VaR are prominent under extreme market conditions.

In this case, for the dynamic margin system of the Bitcoin futures market, it may as well apply the risk measurement of Bitcoin pricing to the setting of the futures maintenance margin. For example, the greater the variance (risk), the more unstable the current Bitcoin pricing and the tendency of market volatility to increase. Therefore, the futures maintenance margin should be increased to better compensate for the risk of contract price fluctuations. The specific details of this application scenario need to be further discussed in the future.

Summary

This article discusses what kind of data is used to measure the risk of bitcoin pricing, and believes that the buy 1 and sell 1 price data of major spot exchanges should be used. Furthermore, we use variance to describe the degree of dispersion between the bid 1 price and the sell 1 price, and describe the risk of current Bitcoin pricing.

The risk measurement of Bitcoin pricing can be applied to high-frequency quantitative strategies and the setting of futures margin. Two examples are given to illustrate:

In the high-frequency quantification strategy, the greater the variance (risk), the more unstable the current Bitcoin price will be. The next moment (the moment we place the order at the exchange), the degree of price fluctuation will increase. , Which may make the originally profitable operation become unprofitable. Therefore, this variance (risk) can be used as an auxiliary indicator to participate in the risk control of high-frequency strategies.

In the setting of bitcoin futures margin, for the dynamic margin system, the greater the variance (risk), the more unstable the current bitcoin pricing, and the market volatility tends to increase. Therefore, the futures maintenance margin should be increased. In order to better compensate for the risks caused by contract price fluctuations. The specific details of this application scenario need to be further discussed in the future.

References

[1] Ding Jingyuan. Research on Financial Complex System Modeling and Dynamics Mechanism. Shanghai University Doctoral Dissertation, 2011. 57

[2] Yang Jiaojiao. Research on the Construction and Application of Maintenance Margin Setting Model for Futures Trading. Master’s Degree Thesis of Hunan University, 2014. 1-2

[3] Wen Wen. Empirical Research on Margin Design of China Commodity Futures Market. Master’s Thesis of Tianjin University, 2012. 3

Publisher: XBITRUST & Paiclub Capital

Author: Su Wenjie

 

Posted by:CoinYuppie,Reprinted with attribution to:https://coinyuppie.com/risk-measurement-of-bitcoin-pricing/
Coinyuppie is an open information publishing platform, all information provided is not related to the views and positions of coinyuppie, and does not constitute any investment and financial advice. Users are expected to carefully screen and prevent risks.

Leave a Reply