產品相似性與信用傳染

Abstract

在本篇論文中，透過Hoberg and Phillips (2010)分析10-K報表上產品資訊，藉由文字算出各家公司的關係度，找出產品接近的公司，做兩兩配對，透過此種方式可以更即時的反應公司的現況，也使用了產業變數、會計變數、市場變數以及總體經濟變數，試圖使用這些變數測量產業內的相互關係，產業變數為使用產品資訊後，所定義的同產業之公司所計算出的違約距離平均數，此外當有公司破產時在當月份認定為同產業的公司產生最大的影響隨後遞減，本篇論文主要模型為Duan et al (2012)遠期違約的強度模型，透過此模型了解產業內的相互關係，對於違約機率是否有影響。經由實證結果可以發現不論在產業變數、市場變數以及總體經濟變數都對於破產機率有顯著的效果，也發現了產業的競爭效應，也可以藉此說明產業內的信用傳染效應。

This paper examined the intra-industry interactions using the method of Hoberg and Phillips (2010). Hoberg and Phillips (2010) used the text-based analysis of 10-K product description and computed the correlation of companies, finding each pair of firms with relevant products. It can efficiently and instantly reflect the companies’ status in this way. We apply different specific variables such as the industry variable, which were used in Hoberg and Phillips (2010), to calculate the mean of distance to default. In this paper, we also use the forward intensity approach of Duan et al (2012) to find out whether the variables we select can affect the default intensity. Finally, the empirical evidence in this paper shows the intra-industry competitions and the contagion effect.