國際市場資產風險與波動性分析

Abstract

本文透過時間數列分析和變異數異質性模式(GARCH)，探討全球資產波動性風險與波動性分析。研究涵蓋四個市場大盤指數、四個著名科技股、一個加密貨幣、兩個外匯市場及原物料資產共分析十三種資產，分析其在疫情前、疫情期間和疫情後的波動特徵。本研究根據時勢挑選出特定資產與股票指數，透過R套件抓取Yahoo Finance擬研究的資產，計算夏普值與日報酬率，月報酬率。並通過Ljung-Box Q與ARCH-LM test判斷是否有ARCH效應，決定是否採用GARCH模型，並根據數據型態配適合適的GARCH模型估計指數的變異與波動，其中透過不對稱檢定Joint test檢定非對稱性，如果具有非對稱性則採用GJR-GARCH模型。本研究擷取四段區間進行分析探討，分別為完整週期、疫情流行前、疫情期間與疫情流行期間後。2017年至2019年被定義為疫情流行前， 2020年1月1日至2022年8月31日為COVID-19盛行期間， 2022年9月至2024年9月則探討疫情後的股市波動。結果顯示在疫情期間造成眾多股市動盪，高風險資產如輝達和比特幣的波動性顯著，避險資產如黃金和原油在市場不確定性下波動加劇。GARCH和GJR-GARCH模型能有效捕捉資產報酬率的波動特徵，並通過樣本外一步預測方法提高預測準確性。在樣本外預測結果顯示，波動最大的為輝達，在疫情後的波動顯著，說明了高風險資產的特性。納斯達克綜合指數和巴黎CAC指數的波動性反映出市場與經濟形勢的影響。黃金期貨在市場起伏不定時的避險需求增加。日元對美元匯率和歐元對美元匯率的波動顯示出各國貨幣政策的影響。2022年初疫情影響逐漸消散後，俄烏戰爭使各資產波動率大幅上升，尤以布萊特原油影響最為顯著。2024年，布萊特原油的波動率再次回升，可能受到OPEC+減產政策與能源需求變化的影響。

This study employs time series analysis and Generalized Autoregressive Conditional Heteroskedasticity (GARCH) models to explore the volatility and risk characteristics of global assets. The analysis encompasses 13 assets, including four market indices, four prominent technology stocks, one cryptocurrency, two foreign exchange markets, and commodity assets, across three periods: pre-pandemic, during the COVID-19 pandemic, and post-pandemic.Data for the selected assets were obtained from Yahoo Finance using R packages, with Sharpe ratios, daily returns, and monthly returns calculated. Ljung-Box Q and ARCH-LM tests were applied to detect ARCH effects and determine the suitability of GARCH models. Asymmetric effects were examined using the Joint test, with GJR-GARCH models employed for assets exhibiting asymmetry. The study analyzed four intervals: full period, pre-pandemic (2017–2019), pandemic (January 2020–August 2022), and post-pandemic (September 2022–September 2024). The findings indicate significant market volatility during the pandemic. High-risk assets, such as NVIDIA and Bitcoin, exhibited pronounced volatility, while safe-haven assets like gold and crude oil experienced heightened fluctuations amid market uncertainty. GARCH and GJR-GARCH models effectively captured asset return volatility, with out-of-sample one-step-ahead forecasts enhancing predictive accuracy. NVIDIA showed the highest volatility, especially post-pandemic, reflecting the inherent risks of high-risk assets. The volatility of indices such as the NASDAQ Composite and CAC 40 revealed the influence of market and economic conditions. Gold futures saw increased demand during market turbulence, while currency pair volatilities (e.g., JPY/USD and EUR/USD) reflected the impacts of monetary policies. After the pandemic subsided in early 2022, the Russia-Ukraine war caused a surge in asset volatilities, most notably in Brent crude oil. In 2024, crude oil volatility rose again, likely driven by OPEC+ production cuts and changes in energy demand.