探索台灣可轉換公司債發行條款之設計

Abstract

可轉換公司債是台灣上市公司籌資的重要管道之一，但是由於可轉債組合有許多權利，對於可轉債的價值以及對於公司股東、債權人等的影響較為複雜。本研究探討可轉債重要發行條款設計的影響因素以及重要發行條款間在設計上是否有關連及其關連程度。 研究首先探討可轉債在台灣發行的現況以及重要發行條款對於可轉債價值的可能影響，進而透過文獻回顧，瞭解可轉債的發行動機、發行公司的特性以及可轉債的發行效果，初步瞭解可轉債的發行的經濟意義以及設計可轉債的重要影響因素。 繼而引進證券設計理論，在證券設計的理論架構下，可以透過可轉債的發行條款的設計，減輕代理問題，並在資訊不對稱的情況下，作為描述公司特性的訊號。因此我們擬針對可轉債的發行條款設定代理變數進行研究，我們選取A.轉換後股權比例、B.發行期間、C.相對禁止贖回期間作為模式中的內生變數。 考慮研究變數間的內生關係，我們利用聯立系統方法中的三階段最小平方法(3SLS)進行探索，研究結果顯示，公司規模、財務槓桿、信用品質、異常報酬、董監持股比率等變數對於可轉債的設計有顯著影響性存在。我們推測公司可以透過發行條款的設計，將公司經營的風險以及未來盈餘成長性等資訊，發射到市場中，並將風險移轉給投資人，適度的反映了代理問題的存在。若不利用系統方法，使用普通最小平方法(OLS)亦顯示相同的關係存在。 而在設計條款的相互關係上，研究顯示其關連性不高，意味著公司設計可轉債時，乃以實際資金需求作規劃，發行條款間的交互關係並非其關心的重點。

Financing by convertible bonds (CB) is one of the important way to get money. But for the complexity of the convertible, to estimate the value of CB and to measure the effect of issuance is very difficult. We will study the factors of designing CB's important clauses. At the first, we get know about the status quo of CBs at Taiwan. And we estimate the influence of designing clauses to CB's value. Then, by articles review, we understand the cause of issue CB, the characteristic of the firms, and the effect of issue CB. On the theorical structure of security design, the agency problem may be eased by proper designing bond clauses. And at the condition of information asymmetric, we can also get know about characteristic of the firm by the signaling effect of CB. To study the CB's clauses, we use A. post-conversion equity ownership, B. maturity, C. relative length of call period protection as the endogenous variables of the model. We use three stages least square method (3SLS) to explore the endogenous relationship of the three variables and the important factors of CB design. It shows that A. firm size, B. financial leverage, C. credit quality, D. abnormal earning, E. stocks proportion holding by directorate will have significant influence to CB design. We argue that CB issuers signal the business risk and the growth opportunity to the market and ease agency problem by designing CB's clauses. We will get the same result by using ordinary least square method (OLS). But we find there is no significant relation among the three endogenous variables. It means that the endogenous relation among CB clauses is not the emphasis when designing CB.