比特幣匯率之實證分析

Abstract

比特幣創始初期的匯率一直維持相當平穩趨勢，直到 2011 年 2 月成交匯率才首度突破 1 美元，然近年來其成交匯率波動加劇，2017 年 11 月 8 日收盤價已來到7,458.79 美元，7 年間投資報酬率高達 280 萬倍，可稱為本世紀漲幅最高之投資標的。本研究透過文獻探討並對比特幣的供給與需求面進行分析，探討影響比特幣匯率波動的因子。 本文將比特幣匯率函數以黃金價格、北海布蘭特原油期貨價格、西德州原油期貨價格、美元指數、日圓匯率、人民幣匯率及恐慌指數作為解釋變數，使用資料樣本期間為 2013 年 12 月至 2017 年 6 月，將研究期間內沒有共同交易日之日資料刪除，每個變數觀測值各有日資料 845 筆。首先將各變數資料取自然對數後進行單根檢定，非定態之變數則再經一階差分後再次檢定是否為定態時間序列資料，建立變數間向量自我迴歸模型(Vector Autoregression model,VAR)觀察比特幣匯率與各變數落後期之相關性，由衝擊反應函數了解外生衝擊對比特幣匯率的影響，並以 Granger 因果關係檢定變數之間是否存在領先、落後或獨立關係，最後將向量自我迴歸模型中與比特幣當期匯率呈現顯著之變數作為多元迴歸模型的解釋變數，建立多元迴歸模型探討影響比特幣匯率的因素。研究結果顯示，落後一期的日圓匯率對當期比特幣匯率影響最大且為同向關係，落後一期的黃金價格與當期比特幣匯率為同向關係，同時落後一期的比特幣匯率與當期比特幣匯率為負向關係。 雖然模型當中此三變數的係數值均為顯著，但解釋能力偏低，表示研究期間比特幣匯率變化的隨機性，比特幣創始來源與經濟因素幾無關聯性，獨立於金融體系的任何資產，是以其匯率巨幅震盪可能與人為炒作或政策因素有關。

The initial exchange rate of bitcoin has maintained a fairly steady trend, it exceeded $ 1 for the first time in February 2011. However, the exchange rate volatility in recent years. The closing price of November 8, 2017 has reached $7,458.79. During this period, the return on investment of bitcoin as high as 280,000% , it can be called the most popular investment target in this century. This thesis explores and analyzes the supply and demand of bitcoin through literature to explore the factors affecting the exchange rate of bitcoin. To figure out the factors, we select gold price, oil prices, US Dollar Index, JPY against USD exchange rate, CNY against USD exchange rate and VIX as explanatory variables. From December 2013 to June 2017, there are 845 observations. First, we use unit root test to ensure all the series data used in regression analysis are stationary. Then, we create the VAR model to analyze how explanatory variables affect the exchange rate of bitcoin and use impulse response function to figure out the response of bitcoin to the exogenous shock of other variables. Granger causality test is also used to determine whether a time series is useful in forecasting another. Eventually, we use the statistically significant variables in multiple regression analysis to investigate the impact factors for the exchange rate of bitcoin. The conclusion of this research shows that the exchange rate of bitcoin is affected by first lag of itself, first lag of gold price and first lag of JPY against USD exchange rate. Although the coefficient of these three variables in the model all significant. However, the low explanatory ability indicates the randomness of changes in bitcoin exchange rates during the study period. Due to bitcoin is an independent asset, the exchange rate volatility may be a result of manipulation or policy.