本文研究比特幣報酬與風險之間的關係，並以Corsi (2009) 提出之異質自我迴歸已實現波動率 (Heterogeneous Autoregressive Model of Realized Volatility, HAR-RV) 模型計算波動率，透過引入具有長記憶 (long memory)特性的條件變異估計式，來分析加密貨幣市場之風險報酬抵換關係。使用2018年1月至2022年9月期間之比特幣的五分鐘日內高頻資料，同時考慮上、下行風險以及COVID-19流行期間造成金融市場動盪之影響，實證結果顯示比特幣的風險與報酬之間存在顯著的正向關係，同時投資人在面對下行風險時會要求更高的報酬，此現象在COVID-19流行期間尤其明顯。;This thesis studies the risk-return relationship in the cryptocurrency market and uses the Heterogeneous Autoregressive Model of Realized Volatility (HAR-RV) proposed by Corsi (2009) to calculate the volatility. The HAR-RV model is a conditional variance estimator with the feature of long memory. Using the five-minute intraday high-frequency data of Bitcoin from January 2018 to September 2022, we measure upside and downside risks and take the impact of COVID-19 into account. The empirical results provide evidence of a significant positive relationship between the risk and return of Bitcoin, and investors incline to require higher returns when facing downside risks, especially during the COVID-19 epidemic.
