本文藉由Johansen共整合向量與Granger因果關係檢定探討波羅的海巴拿馬極限型船運價指 數（Baltic Exchange Panamax Indices; BPI）與巴拿馬極限型船船價之因果關係。本研究實 證後得到以下幾點結論：(1)單根檢定顯示指數與船價之時間序列，呈現非定態序列 (Non-stationary Series)；經一次差分後指數與船價之時間序列，呈現定態序列(Stationary Series)，有助於增加本研究之精確性。(2) Granger因果關係檢定，顯示BPI與船價間具有單 向因果關係。(3)Johansen共整合檢定，顯示BPI與船價存在長期之均衡趨勢。(4)變異數分 解與衝擊反應函數檢定結果，驗證二手船價在短期受BPI與自身之影響較大；但長期而言， BPI與新船之船價則對二手船價有較大之解釋能力；另外，BPI與二手船價對新船船價之影 響有限。另外進行季節效應檢定發現巴拿馬極限型船運價指數有顯著之季節效應，但船價 並無明顯之季節性變動。

The methodology of this paper adopted Vector Auto-regression approaches to test the causality between the Baltic Panamax Indices and ship price. The purpose of this paper could provide the characteristic and information transmission situation of BPI and ship price, but also utilize the effective reference information to draw up investment portfolio strategy. The finding on this paper could be summarized as: (1) empirical tests of BPI and ship are non-stationary. However, they are all stationary after using first difference. （2）The ship price is Granger Cause by BPI. (3) Johansen co-integration test is discovered that the ship price and BPI are cointegrated with one common stochastic trend. (4) Adopting impulse response analysis and variance decomposition approaches the secondhand ship price impulse was responded by owned past variance and BPI at short strength that was impulse responded by BPI and new-build ship price at long strength. Moreover, the BPI have presented relative high level or low level on the specific month every year, but ship price hasn't a significantly seasonal effect apparently.