【作者】 李国荣；

【导师】 吴大为；

【作者基本信息】 大连理工大学 ， 应用数学， 2006， 硕士

【摘要】 现有的大多数关于违约传染的文献都假设两个负债人之间的违约是一个直接的因果关系。Sch(?)nbucher于2003年提出信息影响也可导致违约传染,并给出了一个仅有信息影响而无因果联系的违约传染模型。在此模型中,投资者仅有负债人违约风险真正大小的不完全信息,而真正的违约风险依赖于一系列不可观测的随机变量,并且这些变量在负债人之间是相关的。但是,它们却难以估计和量化,一般都带有误差项。于是Sch(?)nbucher(2003)利用比例危险模型,令h_i=Y_iλ_i,其中,Y为连续、非负的随机变量。 本文推广了Sch(?)nbucher(2003)模型,假设负债人的违约时间T_i服从Г(κ,β_i)分布,同时假设尺度参数β_i=Y_iλ_i,并为了具有传染效应,假设它们相等,在不同信息条件下,具体给出了负债人不同的生存概率与违约危险率。最后,本文对首次违约债券进行定价。更多还原

【Abstract】 Much of the existing literature on dafault contagion assume a direct causal relationships between two obligors’ defaults. Schonbucher(2003) present a model in which default contagion arises without causal links solely from information effects. In this model, investors have only incomplete information about the true size of any obligors’ risk of default. The true default risk depends on a number of variables which none of the market participants can observe directly. Furthermore, many of these hidden variables maybe correlated across obligors. However, they are notoriously difficult to estimate and quantify, almost all of them carry large estimation errors. Thus, Schonbucher(2003) use proportional hazards model asuming that the default hazard rates are not fully observable but they carry a multiplicative error term Y, i.e. h_i = Y_iλ_i. where Y is a continuous nonnegative random variable.This paper present an extension of the Schonbucher(2003) model, with the assumption that the default time obeying Γ-distribution and the scale parameter β_i = Y_iλ_i. For exhibating contagion properties, we assume all Y_i are equal. Under different information , the obligors’ survival porbabilities and default hazard rates are discussed. Finally, we price the first default bonds.更多还原