Bank loan pricing model: an application of ROC curve analysis

• Main Authors: Yu-sen Shih, 施毓森
• Other Authors: Da-bai Shen
• Format: Others
• Language: zh-TW
• Published: 2007
• Online Access: http://ndltd.ncl.edu.tw/handle/51165966407430041091

碩士 === 東吳大學 === 會計學系 === 95 === Bank loan is the most important banking business irrespective of the proportion of the assets or sources of income. However, majority lenders decide lending rate solely based on subjective judgement, but on a systematic and rationalized pricing model. The pressure of competition in bank industry is getting more and more severe over past decade. To maintain the market share or short-term performance of banks, many lenders made mis-pricing high-risk lending, induced high non-performing loan ratio, damaged the interest of stakeholders and depositors and then destroyed the stability of financial system eventually. Accoring to other researchs indicated that there is a mis-pricing problem in Taiwan bank loan market. The thesis trys to introduce a bank loan pricing methodology by integrating Receiver Operating Character curve with specific cost function of lending strategy. Base on the ROC analysis to loan pricing, lenders can decide the optimal cut-off point from the ROC curve and specific cost function of lending strategy. This thesis also indicates how the optimal cut-off point changes when the significant components of cost function change by simulation analysis. A credit spread term structure can be derived from ROC analysis as well and to form a risk-base loan pricing model. According this thesis, the most important function of the ROC analysis for loan pricing are: 1. to determine the risk-based optimal cut-off point and ensure minimizing the Type I and Type II error. 2. lenders can alter the optimal cut-off point by adjusting the component of cost function of lending strategy or changing the value of some variables dynamically, to sustain the incoming economic situations. 3. to establish a credit spread term structure then set discriminating interest rate. Lenders with more powerful rating system will have lower Type I and Type II error and will make more effective price than those who use weaker rating system.