公司之平均股票報酬是否可由貝他外之基本面變數協助解釋或預測，是最近財務實證研究重心之一。本益筆、淨值市價筆、現金流量筆、及公司規模是常被發現有此能力之變數，而Bhandari ( 1988 ) 曾提出證據說明融資變數也是其中之一。 本文認為融資是公司主要決策之一，不應和本益比等外生變數一起放在迴歸式右手邊分析；不同融資順位之公司可能有不同之財務風險，會使股票報酬與基本面之關係作結構上之轉移。因此本文採取與以前文獻不同之方法，並不將融資變數直接置於迴歸式之右手邊，而你採用Ordered Probit Selection模式分二階段處理。第一階段將公司依融資順位型態分為四群，及內部融資、銀行貸款、發行可轉換公司證券及現金增資。第二階段則依不同之順位群，分別分析其股票報酬率與基本面變數間之關係。 我們利用民國七十六年到八十一年間之台灣上市公司之財務資料分析，結果發現，不同之族群的確發現不同之顯著基本面變數來解釋股票報酬。淨值市價比變數是發行新股票群之顯著變數，現金流量比則屬內部融資群，公司規模則在大部份之群皆顯著，但本益比則在任一群皆不顯著。

We classified firms' financing decisions into four categories: internal financing banking, bank loans, convertible bonds and preferred stocks, and new common shares. We use the ordered probit selection model to investigate the relation between stock re-turns and market fundamentals for each category. At the first stage, the probabilities of selecting one of the four financing decisions are estimated. Then, for each choice catego-ry, we examine the relations between stock returns and market fundamental ariables; E/P, B/M CF and size, given market betas. The empirical results using financial data of listed firms in Taiwan show the im-portance of classifying firms into different categories. We find that different category of firms may find different significant explanatory variabkes to financing decisions, so do different fundamental variables to stock returns. This paper uses pooled cross-sction and time series data. We adjust fudamental variables to account for changes in stock prices overtime. When the adjusted data is used, we find that BIM has explanatory power to stock returns for the new-issue catego-ry; CF for the internal-financing category. The sign of both are positive. These are con-sistent with the findings in other papers. The size variable is also significant in all categories. However, the sign is not consistent. The convertible category has reverse size effect and others have size effects. As for the EIP variable, we don't find any additional explanatory power to stock returns for any category. The use of different definition of size hardly changes our results. However, the size variable is the most consistent and persistent explanatory variable to stock returns given market betas.