為研究企業資本結構之適用性，Shyam-Sunder and Myers(1999)以簡單線性 迴歸模型來驗證靜態抵換理論(Static Tradeoff Model; Myers,1977)及融資順位理 論(Pecking Order Model; Myers and Majluf,1984)二種學說。依其實證分析結果， 認為融資順位理論對美國企業融資行為的解釋能力優於靜態抵換理論。但是， Chirinko and Singha(2000)對其實證結果提出質疑，他們認為若僅依線性融資順 位模型來判斷企業資本結構，在實證上可能存在誤謬。為解決這些問題，本研 究嘗試提出二個有拗折點的融資順位模型，以線性迴歸模型及本文所提出的二 種修正模型，來修正Shyam-Sunder and Myers的模型，以解決Chirinko and Singha 所指出的誤謬。並以台灣336家非金融業的上市公司為研究對象，探究台灣產業 之資本結構，並驗證模型的適用性。

In order to study capital structure decisions, Shyam-Sunder and Myers(1999) introduced a simple linear regression model to test the Static Tradeoff Model (Myers,1977) and the Pecking Order Model (Myers and Majluf,1984). According to their study, the Pecking Order Model is better than the Static Tradeoff Model in explaining companies’ financing behavior. However, Chirinko and Singha(2000) claimed that there are two problems happened in Shyam-Sunder and Myers’s model. This paper intends to propose another two Pecking Order Models with bending point that modify the linear regression model and solve these mistakes found by Chirinko and Singha. Furthermore, we will use 336 non-financial companies that listed in Taiwan Securities Exchange to test the three models. We find pairing contrast test among the three models can clear up the mistakes that claimed by Chirinko and Singha.