股票市場及美元指數與原油在俄烏戰爭前後之相關性探討

本研究以原油價格、股票市場及美元指數對為研究標的，樣本資料區間取自2021年2月24日至2023年2月24日,以非綫性門檻誤差修正模型為架構，分別在俄烏戰爭事件前後探討原油價格對股票市場、美元指數之長短期因果關係，利用動差門檻自我回歸模型（M-TAR）進行門檻共整合檢定並進一步采用Enders and Granger（1998）及 Enders and Siklos（2001）門檻誤差修正模型（TECM），來詳細探討該變數間的長短期非對稱因果關係。　　實證結果指出，第一：俄烏戰爭前不論是美元指數或標普指數與布倫特原油皆存在長期非綫性共整合關係,標普指數與布倫特原油互動關係更進一步存在非對稱關係；俄烏戰爭後，僅美元指數與布倫特原油存在長期非綫性門檻共整合關係，標普指數與布倫特原油間在戰後不存在共整合關係。第二：在短期因果關係上，標普指數或美元指數與布倫特原油短期互動關係在俄烏戰爭前、後並無發生改變，皆不存在短期因果關係。第三：在長期因果關係上，俄烏戰爭發生前美元指數或標普指數與布倫特原油間不存在長期因果關係；俄烏戰爭發生後，僅布倫特原油對美元指數存在長期因果關係，當位於門檻值之上時，布倫特原油對美元指數存在領先且顯著之單向「領先-落後」關係。本研究藉由實證結果分析得出，俄烏戰爭作爲重大地緣政治事件，其發生影響股票市場及美元指數與原油之相關性，原油與股票市場在戰後存在相關性減弱現象，原油與美元指數在戰後相關性增強，短期內對美元指數具有一定之影響。

關鍵字

布倫特原油；俄烏戰爭；門檻誤差修正模型；門檻共整合模型

並列摘要

This study examines the causal relationships between crude oil prices, stock markets, and the US dollar index using a nonlinear threshold error correction model. The sample period spans from February 24, 2021, to February 24, 2023, focusing on the period before and after the Russia-Ukraine war. The study employs a threshold autoregressive model (M-TAR) to test for threshold cointegration and further applies the threshold error correction model (TECM) proposed by Enders and Granger (1998) and Enders and Siklos (2001) to investigate the long- and short-term asymmetric causal relationships among the variables. 　　The empirical results indicate the following: First, before the Russia-Ukraine war, both the US dollar index and the S P 500 index show long-term nonlinear cointegration with Brent crude oil. Furthermore, there exists an asymmetric relationship between the S P 500 index and Brent crude oil. After the Russia-Ukraine war, only the US dollar index exhibits long-term nonlinear threshold cointegration with Brent crude oil, while no cointegration is found between the S P 500 index and Brent crude oil. Second, in terms of short-term causal relationships, there are no significant changes in the short-term interactions between the S P 500 index or the US dollar index and Brent crude oil before and after the war, indicating no short-term causality. Third, regarding long-term causality, there is no long-term causal relationship between the US dollar index or the S P 500 index and Brent crude oil before the Russia-Ukraine war. However, after the war, Brent crude oil exhibits a long-term causal relationship with the US dollar index, with significant one-way "leading-lagging" causality when it exceeds the threshold value. Based on the empirical findings, this study concludes that the occurrence of the Russia-Ukraine war as a significant geopolitical event affects the correlation between stock markets, the US dollar index, and crude oil. The correlation between crude oil and stock markets weakens after the war, while the correlation between crude oil and the US dollar index strengthens. Crude oil has a certain impact on the US dollar index in the short term.