利用迴歸樹分析台灣航空和航運類股報酬率的影響因素

本研究主要目的在於找出影響台灣航空及航運類股超額報酬的因素。樣本資料取自台灣經濟新報（TEJ），共計38間上市公司，研究期間為西元2000年第一季至西元2019年第四季，共計76季。本文將20多個財務報表變數視為自變數，並且為了探討股票報酬率與總體經濟的關係，加入了對數西德州油價、對數布蘭特油價、美國經濟政策不確定性指標、美國報紙文字爬蟲不確定性指標、對數美國S&P 500指數。利用R語言決策樹中CART演算法檢視影響台灣航空及航運類股報酬率的影響因子。結論如下：第一、週轉率對整體樣本的影響是最大的，其次為現金流量比率以及總負債/淨值；第二、總體經濟的波動對於航空及航運業的影響不顯著；第三、決策樹的分類越詳細，產生的階層數越多，所產生的超額報酬越小。

關鍵字

超額報酬；決策樹；台股；航空業；航運業

並列摘要

The main purpose of this research is to find out the factors that affect the excess returns of Taiwanese airline and marine shipping stocks. The data is taken from the Taiwan Economic Journal (TEJ). Thirty eight listed companies are included. The research period is from the first quarter of 2000 to the fourth quarter of 2019, a total of 76 observations per company. This article uses more than 20 financial statement variables as independent variables. In order to explore the relation between stock returns and the macro-economy, we include the log of West Texas oil price, the log of Brent oil price, U.S. Monthly Economic Policy Uncertainty (EPU) Index, the index of U.S. News Coverage about Policy-related Economic Uncertainty, and the log of S&P 500 index as additional explanatory variables. Decision tree is constructed using the R statistical software program. The conclusions are as follows. Turnover rate has the largest impact on excess return, followed by the cash flow ratio and debt-to-equity ratio. Second, macro-economic fluctuations have a little impact on the return of Taiwanese airline and marine shipping companies. Third, the more complex the classification of the decision tree, the smaller the excess return observed.