| 摘要: | 本研究利用時間序列方法，包括單根檢定、共整合檢定、Granger因果關係檢定、鄒檢定、向量自我迴歸模型、向量誤差修正模型，以探討燃油價格與海運運價及海運運價彼此間的關聯性。其中，海運燃油價格以IFO380燃油價格指數，海運運價則分別包括：上海出口集裝箱運價指數(SCFI)、波羅的海乾散貨運價指數(BDI)及波羅的海油輪運價指數(BDTI)。研究期間為2017年1月1日至2022年12月31日共六年之週資料。
實證分析結果發現：(1) 所有研究變數皆為非定態(nonstationary)，須經過一階差分才能呈現定態。(2) 透過Johansen共整合檢定，僅有SCFI與其他運價指數與燃油指數在結構轉變時間點後未有共整合關係，其餘運價指數與燃油指數皆具有共整合關係。(3) 經過鄒檢定發現序列皆具有結構轉變，其中：IFO380燃油價格指數的結構轉變時間點在2020年3月9日，SCFI的結構轉變時間點在2020年8月31日，BDTI的結構轉變時間點在2020年6月1日，BDI的結構轉變時間點在2020年6月15日。(4) 我們利用VECM與VAR模型分別在結構轉變點前、後，來捕捉各變數之間的短期關係，發現在結構轉變點前後變數間的短期關係有明顯的變化，我們因此推測重大事件的發生，可能會改變變數間的短期關係。(5) 各航運市場在結構轉變時間點前後的Granger因果關係如下：在貨櫃市場中，不論在結構轉變時間點之前、後，貨櫃運價指數SCFI皆會領先於燃油價格指數IFO380與BDTI指數。在散裝市場中，其在結構轉變時間點前，散裝運價指數BDI領先於IFO380指數。然而在結構轉變時間點後，BDI指數與IFO380指數則無granger因果關係。在油輪市場中，其在結構轉變時間點前，IFO380指數領先於油輪運價指數BDTI。而在結構轉變時間點後，則轉為BDTI指數領先於IFO380指數。
本研究希冀提供航商、船東及投資者透過追蹤燃油價格與航運價格及航運價格彼此之間的關係，以提前佈局船隊經營及達成降低投資風險的目標。
;This study uses time series methods, including unit root test, cointegration test, granger causality, chow test, vector autoregressive model, and vector error correction model to explore the interaction between bunker fuel prices and freight rates. We selected IFO380 as the shipping bunker prices, and the shipping price index consists of three major freight indexes: Shanghai Containerized Freight Index (SCFI), Baltic Dry Index (BDI), Baltic Dirty Tanker Index (BDTI). The study period was from January 1, 2017, to December 31 2022 week of data.
The results of the empirical analysis found that: 1. the variables are non-stationary state, rendered after a first-order differential steady state. 2. Through Johansen′s cointegration test, only SCFI is not co-integrated with other freight rates and bunker indices after the structural break. In contrast, all other freight rates and bunker indices are co-integrated. 3. The Chow test shows that all the variables have structural breaks. The structural break date of IFO380 is on March 9, 2020; SCFI is on August 31, 2020; BDTI is on June 1, 2020; and BDI is on June 15, 2020. 4. We use the VECM and VAR models to capture the short-term relationships between variables before and after the structural break point and find significant changes in the short-term relationships between variables before and after the structural break point. Therefore, we hypothesize that the occurrence of a major event may change the short-term relationship between variables. 5. The Granger causality test for each shipping market before and after the structural breakpoint is as follows: In the container market, the SCFI leads the IFO380 and BDTI before and after the structural break point. In the bulk market, BDI leads IFO380 before the structural break point. However, there is no granger causality between BDI and IFO380 after the structural break point. In the tanker market, IFO380 leads BDTI before the structural break point, while BDTI leads IFO380 after the structural break point.
This study aims to allow carriers, shipowners, and investors to track the relationship between bunker prices and shipping prices to plan their fleet operations and achieve the goal of reducing investment risk. |
