油價與原物料價格對海運運費波動預測之影響

The effects of oil and commodities prices on dry-bulk shipping freight rate volatility forecasts

陳主宜(Jhu-Yi Chen)

指導教授 ： 邱建良 ； 劉洪鈞

淡江大學/商管學院/財務金融學系碩士班/碩士(2013年)

https://doi.org/10.6846/TKU.2013.00922

摘要

波羅的海指數(BDI)代表散裝船運業的景氣指標，可作為投資人欲投資海運類股之重要參考依據。因此，本研究以BDI指數作為被解釋變數、西德州中級原油價格(WTI)與商品期貨價格指數(CRB)作為解釋變數，利用時間序列的方法探討BDI指數的波動度特性。分別使用GARCH、IGARCH、GJR-GARCH與QGARCH等波動度模型進行配適。此外，利用前述模型，根據移動式窗估計法預測BDI指數之週波動性後，再以 MAE、MSE、RMSE、LL、MME(O)、MME(U)等損失函數，進行樣本外波動預測能力評估。根據前述四種模型進行樣本內估計與樣本外預測，具體實證結果如下：(1)由GARCH類模型之平均數方程式的參數估計結果發現，BDI指數報酬率不僅受到歷史資訊的影響，亦受油價衝擊使得航運類股之獲利情形有負面的影響；商品期貨價格對航運類股獲利情形有正面的影響。(2)由變異數方程式的估計結果得知，BDI指數之波動性明顯受到歷史波動性的影響；前一期原油報酬率對於BDI指數波動性為顯著的負向影響；然而，BDI指數波動性受到前一期商品期貨價格指數報酬率的影響並不顯著。(3) BDI指數報酬率存在顯著的波動持續性及不對稱的槓桿效果。(4)若以報酬率的平方作為波動度的代理變數，IGARCH模型的樣本外波動預測能力優於其他模型；若改以價格變幅作為波動度的代理變數，則以GARCH模型的表現最佳。以上實證結果可提供相關海運業者及股市投資人從事避險操作與擬訂投資決策之重要參考依據。

關鍵字

BDI指數 ； 原油 ； CRB指數 ； GARCH

並列摘要

Baltic dry index (BDI) represents the booming target of bulk ship transport industry. It’s can be a reference for investment who want to invest in the maritime sector. Therefore, this thesis adopts the time series method to explore the nature of BDI index volatility. BDI represents explanatory variable, while West Texas intermediate (WTI) oil volatility and commodity research bureau futures (CRB) price returns represent dependent variables. We use four competitors, including the GARCH, IGARCH, GJR-GARCH, and QGARCH models, to compare their in-sample goodness of fit and out-of-sample volatility forecasting ability under oil and CRB shocks. Under the aforesaid econometric methodologies, a number of solid evidence has emerged from this thesis. First, the parameters of mean equations in the GARCH genre of models indicate that the BDI index returns are influenced by historical information, while the oil (commodity futures) price has a negative (positive) impact on the profits of shipping stocks. Second, the model estimates of the variance equations in the GARCH families reveal that volatility characteristics of the BDI index are influenced by historical property. Past oil shocks have significantly negative impact on the BDI index volatility, whereas the lag CRB index returns doesn’t significant influences BDI volatility property. Thirdly, high degree of volatility persistence and leverage effect are found in the BDI return series. Finally, out-of-sample volatility forecasting results imply that the IGARCH model provides the most accurate forecasts when the volatility proxy is measured by the squared returns. As for the forecasting results obtained from the price range (PK), the GARCH model is superior to the other competing models. The empirical findings provide crucial implications for marine transportation entrepreneur and common investors to improve decision makings and hence avoid the market risk.

並列關鍵字

BDI ； Crude oil ； CRB index ； GARCH

參考文獻

7.陳永順(2008)，國際散裝乾貨船市場分析，船舶與海運通訊，第十八期，頁12∼29。

15.劉洪鈞、姜淑美、白東岳(2011)，日內資料對黃金指數型基金波動性的增量預測價值，績效與策略研究，第八卷第二期，頁11∼30。

16.蕭堯仁、周恆志(2011) ，波羅地海乾散貨運價指數與金磚四國股價之關聯性，航運季刊，第二十卷第四期，頁1∼24。

3.周恆志、張超琦(2011)，波羅地海乾散貨運價指數之報酬與風險抵換關係，運籌與管理學刊，第十卷第二期，頁31∼39。

1.周明道、林仕展(2010)，波羅的海運價綜合指數與鋼價指數相關性之分析-以VAR 模型之應用，運輸學刊，第二十二卷第二期，頁 211∼232。