應用壓力測試方法評估信用風險

Abstract

近年來在全球化的衝擊下，金融危機發生的頻率逐漸增加，再加上各國陸續實施巴塞爾資本協定(BaselⅡ)，為了能準確預期並有效因應未來可能的金融危機，金融機構極需一個有效之金融風險控管工具。風險值(Value at Risk, VaR)就是一個常被用來評估金融機構信用風險的工具，但風險值無法預測一些極端事件所造成的嚴重損失，而壓力測試(Stress Testing)正好能彌補風險值的缺失。中外關於壓力測試的研究大都是針對全球股市或是金融機構不同的投資組合，使用不同的壓力測試模型來評估市場風險或投資風險，僅有少數的研究是利用壓力測試進行信用風險之研究。故本研究之主要目的是針對金融機構利用壓力測試方法建構一個違約機率與總經變數間的壓力測試模型，本研究先利用相關性分析尋找與違約機率有高度相關之總體經濟變數，再利用逐步廻歸分析(Stepwise Regression)建立一個與違約機率及總體經濟變數相關的壓力測試模型，最後利用壓力情境方法對信用風險進行壓力測試，以有效評估金融機構之信用風險，並提供金融機構在執行各項壓力測試時之參考。本研究最後以台灣某金融機構之實際案例，來說明本研究之有效性。

In recent years, the impact of globalization tends to increase rapidly the occurrences of financial crisis. The financial institutions need an effective and efficient management tool to deal with the financial crisis. Value at risk (VaR) is one of the most popular tools for assessing the credit risk of financial institutions. However, the serious losses caused by extreme events, such as The September 11 terrorist attacks, are unpredictable using VaR, but the stress testing can make up the deficit of VaR. Most of researches focus on developing various stress testing models to assess the market risk or investment risk, but very few studies utilized stress testing to evaluate credit risk. This study presents a stress testing model based on the probability of default and the general economic indicators to accurately assess the credit risk. At the first, correlation analysis is used to select the general economic indicators which have high correlation with the default probability. Stepwise regression analysis is then applied to build the stress testing model. Finally, the credit risk is accessed using the stress testing model with various stress scenarios. The effectiveness of the proposed stress testing model is demonstrated using a real case from a Taiwanese financial institution.