石油股價波動性、資訊不確定性與石油期貨報酬頻率連動性之研究

Frequency Connectedness of Oil-Sensitive Stock Volatility, Information Uncertainty, and Oil Futures Returns

本研究探討石油公司波動連動性的時頻動態是否對石油期貨報酬產生影響。我們認為短期和長期波動連動性分別隱含不同的訊息，並以Baruník and Křehlík（2018）提出的計量模型衡量此一現象。研究結果指出，短期波動連動性與良好訊息相關；長期波動連動性與不良信息訊有關。本研究進一步使用迴歸分析與Hansen（2017）提出的拗折迴歸模型（Regression Kink Model），檢驗波動連動性的時頻動態對後續石油期貨收益的影響。實證結果顯示，投資者往往對良好訊息反應不足，對不良信息反應過度。此結果支持Brown, Harlow and Tinic（1988）提出的不確定性訊息假說。

The paper examines whether the time-frequency dynamics of volatility connectedness for oil companies influences oil futures returns. We hypothesize short- and long-term volatility connectedness contain different information flows. To measure the dynamic volatility connectedness in the frequency domain, the econometric model suggested by Baruník and Křehlík (2018) is employed. Our results reveal short-term volatility connectedness is correlated with good information and long-term volatility connectedness is correlated with bad information. This paper further examines the influence of the time-frequency dynamics of volatility connectedness on subsequent oil futures returns using the ordinary least-squares (OLS) regression model and the regression kink model suggested by Hansen (2017). Evidence from the OLS and regression kink models reveals investors tend to underreact to good information (news) and overreact to bad information (news). The findings support the uncertain information hypothesis proposed by Brown, Harlow and Tinic (1988).