可轉換公司債贖回策略的實證及數值模型比對

Abstract

本文延續林亨利(2012)及黃欣裕(2016)的模型，利用CRR tree建構出公司資產樹，在1.可轉換公司債持有人獨佔持有(monopoly)、塊狀履約(block)以及完全競爭(perfect competitive)三種持有情況之部分轉換策略 2.發行公司在Textbook Call Policy和最大化股東權益(Maximize Equity)兩種贖回策略 3.發行公司會以發行新股或是買回庫藏股來支應轉換的可轉債，對可轉換公司債進行評價，並修正衡量提前贖回可能性的因子，從期望到期日改成條件期望到期日，除去可轉債自願贖回及公司破產的狀況，以利和債券贖回的實證分析相比較。從驗證的文獻實證因子來看，與實證結果吻合的情況如下:在cash flow advantage中，轉換策略competitive且贖回策略為極大化股東權益並以庫藏股支應的模型、在corporate tax rate中，轉換策略block且贖回策略為極大化股東權益的模型及轉換策略為competitive且贖回策略為textbook call policy的模型。另外，在interest中，除了轉換策略competitive且贖回策略為textbook call policy的模型有不顯著的與延遲贖回呈現負相關外，其餘模型皆不明顯；在asset volatility及backdoor equity financing中，除了轉換策略monopoly且支應方式為treasury的模型有不顯著的與延遲贖回呈現負相關外，其餘模型皆不明顯。 最後，在我們的12個可轉換公司債評價模型中，並未取得任何一個模型同時皆符合文獻中實證的結果，因此，我們的模型在實務上的應用可能還是有所受限。

This study continues Lin’s (2012) and Huang’s (2016) studies. By taking advantages of their tree models, we can price convertible bonds (CB) from three aspects. The first aspect is the sequential conversion behaviors including monopoly scenario、perfect competitive scenario and the block conversion. Then, we analyze two different call policies, one minimizes the CB holder’s value (the textbook policy) and the other maximizes the value of equity holders. The final aspect is obtaining the shares for converted bonds by either issuing new equity or by repurchasing outstanding stocks from the market. Besides, we measure the impacts of market factors on early call and estimate this behavior by revised conditional expected maturity that removes the cases of voluntary conversion of convertible bonds and bankruptcy from calculations of expected maturity. Thus it can be compared with the empirical analysis of convertible bonds' call strategies. The impacts of factors form past empirical studies that coincide with numerical analyses from our pricing models are listed as below: for the empirical study of cash flow advantage, the model under the competitive conversion strategy combined with the treasury strategy and maximized equity strategy generate coincided results. For the empirical study of corporate tax rate, the models under the block conversion strategy combined with maximized equity strategy, and under the competitive conversion strategy combined with text call policy generate coincided results. Besides, for analyzing the impacts of the interest level, only the model with competitive conversion strategy combined with text call policy could generate negative correlation with call delay phenomena, other models generate non-significant impacts. For analyzing the impacts of the asset volatility and backdoor equity financing factor, only the model with monopoly conversion strategy combined with the treasury strategy could generate negative correlation with call delay phenomena, other models generate non-significant impacts. Overall speaking, non of twelve convertible bonds’ pricing models generates phenomena that simultaneously coincide with empirical studies collected in this paper. Therefore, the usage of our models may be limited on the practice.