運用在信用風險模型的創新數值方法DFPM

Abstract

信用風險議題越來越受到重視，本文提出一個創新的樹狀數值方法（DFPM）企圖運用在信用風險中的結構型首次通過模型，以離散時間點觀察公司資產與負債門檻值的關係。並嘗試加入償還公司債的跳躍因子及跟隨償還而改變負債門檻以更接近市場資訊。此外，此創新樹狀方法亦可將負債門檻值設定為指數型變動。本文並以因無法償還公司債而宣告違約的茂矽電子公司與博達科技公司作為樣本，觀察償還公司債所面臨的風險是否能提早反應，讓投資者能事先做好防範。

More and more attention focuses on credit risk domain. This thesis suggests a novel lattice numerical method “DFPM” to be applied in first passage model which is structural form models in credit risk. DFPM monitors firm value discretely. And DFPM tries to approach market information by adding asset value jumps due to loan repayment and sets the varying barriers. The thesis studies MOSEL and PROCOMP which defaulted caused by loan repayment .The thesis tries to discover the risk of loan repayment early, and lets investors protect themselves early.