以違約距離衡量美國上市公司之信用風險

Using Distance to Default to Measure the Credit Risk of US Public Companies

吳宇哲(Yu-Che Wu)

指導教授 ： 賀蘭芝

元智大學/管理學院/管理研究所/碩士(2002年)

摘要

衡量信用風險的內部模型尚在初步的階段，目前發展的情形就像是1994年時，市場風險衡量模型RiskMetrics及其他方法如歷史模擬與蒙地卡羅模擬法初被介紹時一樣。不同的信用風險內部模型雖各自有自己的理論與假設基礎，但在非預期損失的設計上卻相當類似，這似乎暗示著隨著理論與模型的發展，最終會出現一如市場風險般，為大眾所接受的模型。而本論文的目的即欲說明信用風險模型為何會於九十年代以來迅速的發展，並就這些新信用風險衡量模型的衡量方法、假設加以介紹，予以比較分析之。另外，本論文擬利用選擇權的評價模式，將公司股權視為一標的物為公司資產，執行價為負債的買權，運用BS模型來估計公司淨值距離違約點的資產標準差個數（即違約距離）來評估美國上市公司之違約機率，以窺此法的優、缺點。
本文的上市公司的市值、股權報酬率資料取自芝加哥大學的Center for Research in Security Prices (CRSP)資料庫；上市公司各年度之會計資料(長、短期負債及資產的會計帳面值)取自於Compustat線上資料庫；上市公司的破產日資料取自於SDC資料庫；而美國一年期的T-bill利率則取自Dtatstream資料庫。本文以違約距離來衡量上市公司的信用風險，並以power curve來比較KMV的違約距離(DD)與Z-score破產預測模型對於識別、區別不同違險風險水準的能力，最後則比較破產公司破產前之DD與S&P的表現。
本研究就破產公司破產前四季進行分析時發現，約77%的破產公司在破產前四季，其違約距離已大幅下降(即違約風險上升)，顯示違約距離(DD)的高低確能反應出公司的信用風險高低；而power curve的研究結果發現：DD的區別風險能力要明顯優於Z-score模型。另外，本研究以DD值與S&P的信用評等相比，發現S&P的rating調整落後於違約距離，表示DD值能較快反映出公司的風險增加。

關鍵字

預期違約頻率 ； 選擇權模式 ； 違約距離 ； 信用風險

並列摘要

The development of internal models for credit risk measurement is at an early stage. Its present level of development is similar to that of market risk in 1994, when RiskMetrics first appeared and was followed by alternative approaches such as historic simulation and Monte Carlo simulation. The goal of this study is to describe why there’s a revolution in credit risk measurement and management recently, and to review the current proposed industry sponsored Credit Value-at-Risk methodologies. Moreover, this study use an option-pricing BSM-type model, viewing the market-value position of equity holders in a borrowing firm as isomorphic to holding a call option on the assets of the firm, to estimate the number of standard deviation move in the asset value from the market net worth to the default point (distance to default) as a credit risk measure with a view to see the strengths and drawbacks of KMV model.
The data of this study such as bankruptcy date, market value of equity, interest rate, and book liabilities comes from SDC, CRSP, Datastream, and Compustat database respectively. This essay uses distance to default to measure credit risk, and test the default measure performance by power curve to assess KMV model''s ability to discriminate between different levels of credit risk. Finally, this study compares the predictive ability between DD and S&P rating.
In conclusion, this study finds out that high-DD companies are high risky ones as expected, and the power curves also show that DD is good at discriminating different levels of risk. As for the compassion of DD and S&P rating, DD is better.

並列關鍵字

KMV ； distance to default ； VaR ； EDF

參考文獻

1.Altman, E.I., “Financial Ratios, Discriminant Analysis and the Prediction of Corporate Bankruptcy.” Journal of Finance, September 1968, pp. 1-10 (Winter).

3.Black, F., and M. Scholes, “The Pricing of Options and Corporate Liabilities.” Journal of Political Economy, May-June 1973, pp. 637-654.

7.Crouhy, M., and R. Mark, “A Comparative Analysis of Current Credit Risk Models.” Journal of Banking & Finance, January 2000, pp. 59-117.

8.Duffie, D., and K. Singleton, “An Econometric Model of the term Structure of Interest-Rate Swap Yields.” Journal of Finance, September 1997, pp. 1287-1321.

9.Geske, R., “The Valuation of Corporation Liabilities as Compound Options.” Journal of Financial and Quantitative Analysis, November 1977, pp. 541-552.