違約風險的估計與衡量－ 以台灣未上市的公開發行公司為例

Abstract

自從1997亞洲金融風暴、2007次貸危機引發的金融海嘯，違約風險的管理也逐漸的受到了各界的重視。本研究主要目的是想借由數據來證實在台灣，未上市上櫃的公司，其違約風險是否可由公開資訊中偵測出來。 以傳統的財務觀念來看，違約風險與預期報酬率因為風險溢酬的關係應呈正相關；但是近年的研究卻發現，在許多市場中，我們無法由預期報酬率直接求得其違約風險，所以在違約風險的計算上相對的沒辦法以單一參數(報酬率)來簡單衡量。 不同於以往違約風險的研究，大多是針對上市上櫃的企業；本研究主要是計算未上市櫃公司的違約風險。2006年公開發行的未上市上櫃公司共779間，其中有702間公司連續三年裡至少每半年提供一次財報，本研究的樣本就取自這702間公司。 由於是未上市上櫃的公司，大多都沒有公定的股價，於是在偵測違約風險上，本研究第一部分使用的是依賴財務報表的Z-Score模型。在做完Z-Score的數據分析後，本研究加入了KMV動態模型與其做比較，但KMV模型需要股價，於是本研究的股價部分乃是取自財報中的 “平均股東權益”來代替，並且確實的由KMV模型計算出違約風險。 Z-Score模型與KMV模型經本研究證實，在計算未上市上櫃的公開發行公司的違約風險時有一定的相關性。更近一步的，我們可以用任一模型有效率的預測一間公司是否具有高違約風險；但不幸的是，我們很難由這兩個模型去推測在台灣任何未上市上櫃的公司，其財務狀況是否安全。

Since Asian Financial Crisis in 1997 and the financial tsunami triggered by subprime mortgage crisis in 2007, default risk management has gradually received its attention from all sectors of the world. The main purpose of this thesis is to confirm whether the default risk of unlisted companies in Taiwan can be detected by using public information. From the view of traditional financial concept, default risk should have a positive correlation with expected returns due to the risk premium; however, recent studies have found that in many markets, we can’t directly obtain default risk from expected rate of return; therefore, it’s hard to measure by single parameter-rate of return when calculating risk of default. Unlike most previous studies of default risk which focus on listed publicly traded companies; this thesis is to calculate default risk for unlisted companies. There are 779 public issued but unlisted companies in Taiwan in 2006, and 702 of which provide their financial reports at least every six months for three consecutive years. The samples of this thesis are from these 702 companies. Because there are no official prices for these unlisted companies, the first part of this thesis is using Z-Score model which depends on financial statements to detect the default risk. After data analysis of Z-Score is done, we add KMV dynamic model to compare the results with Z-Score model. We need the stock price of these companies when we use KMV model, so we use the average shareholder's equity to replace the stock price in this part. In the calculation of default risk for public issued but unlisted companies shows positive correlation between Z-Score model and KMV model. Furthermore, we can use either model to efficiently predict whether a company has a high default risk; but unfortunately, it is difficult to predict whether a company's financial position is secure for Taiwanese public issued but unlisted companies.